crop forecasting system and device

专利摘要:
SUMMARY â € HARVEST PREVISION SYSTEMâ € Limiting, as much as possible, the number of field of research, the present inventory estimates the harvest with high precision, using aerial images and chronological series temporal data in special stage of growth of the agricultural product. A group of parameters is used to determine the evaluation criteria to select the measured agricultural land to be surveyed, using previously accumulated research region images, information about characteristic of the field and aerial images of the research field. Select the measured agricultural land in order to cover as much as possible the dispersion in the series of referred parameters, in addition, select the candidates of the measured agricultural land concentrating in your locality in order to reduce research positions as much as possible. However, to analyze the patterns of the time series of meteorological data in each growth stage, calculating the groups of parameters related by the growth conditions, and to estimate the harvest together with the quantity in (...).
公开号:BR112015031241B1
申请号:R112015031241-1
申请日:2014-05-19
公开日:2020-07-07
发明作者:Yu KITANO；Yoriko Kazama
申请人:Hitachi, Ltd；
IPC主号:

专利说明:

TECHNICAL FIELD
[001] The present invention relates to the system and device that estimate the harvest of agricultural products. TECHNICAL BACKGROUND OF THE INVENTION
[002] Official organizations in Japan and casualty insurance companies in the various countries use accident compensation programs for the main agricultural products, since insurance money is paid depending on the conditions of damage to agricultural land. When paying insurance money, the inspector performs a visual inspection test or a harvest survey for the actual measurement, and calculates the insurance money value by the level of damage.
[003] In addition, a real measurement is also performed outside the damaged agricultural land to obtain the average harvest of the total agricultural land, and another real measurement to estimate in advance a unitary harvest as the agricultural land criterion. These processes are a big cost burden, due to the real rigorous measurements in wide area in the agricultural lands and visual inspection test by the inspectors, besides, there is a problem with the inspectors aging. Also, farmers are very dissatisfied with the deviation of the evaluation result due to the performance of the evaluation per person.
[004] Therefore, a system is required that can centrally forecast all harvests using satellite images that allow the footage of the globe's surface to be shot widely and repeatedly.
[005] The technique in the patent document 1 is recognized as a prior art that estimates the harvest of agricultural products through satellite images.
[006] In the patent document 1, the intensity of backward dispersion in the agricultural land sampled in the first growth phase is shown through the images of the Synthetic Aperture Radar, the method generating the forecast model of rice harvest in which it manages the harvest forecast formula based on the rice growth characteristic, the number of stems, through field research, and the rice harvest forecast method. PRIOR ART DOCUMENT PATENT DOCUMENT PATENT DOCUMENT 1: OFFICIAL PATENT JOURNAL No. 2011-167163 SUMMARY OF INVENTION PROBLEM TO BE SOLVED BY THE INVENTION
[007] A lot of measured agricultural land (according to the data studied) is needed to estimate the high precision when performing the harvest estimate through the spectrum of satellite images. In addition, it takes a lot of people and a lot of time to perform the actual measurement, having even greater impact on cost.
[008] On the other hand, agricultural land as the spectrum is almost different in the aspect of the respective model of the harvest estimate depending on the absence or existence of damage, so there is a problem of degradation of precision when performing the harvest estimate under the same model. Methods to Solve Problems
[009] One of those examples of the invention applied here exists to solve the problem mentioned above and has the following characteristics such as: the reception unit that receives the satellite images including agricultural land; the storage unit that respectively stores the information in ways including the mentioned agricultural land and the location information; the image analysis unit that calculates the quantity of the image characteristic of the agricultural land referred to using the satellite images received including the agricultural land mentioned above with the information of forms and location of the agricultural land referred to; the harvest estimate unit that calculates the expected harvest of the aforementioned agricultural land through the amount of characteristic of the above mentioned image calculated. The aforementioned harvest estimate unit, based on the time series standard of previously stored meteorological data, in each particular growth phase of the agricultural products grown on the aforementioned agricultural land, has the unit of analysis of the time series standards that comes out first group of parameters related to the growth condition of the agricultural products mentioned above.
[0010] This unit calculates the expected harvest of the agricultural land mentioned above through the number of characteristics of the image mentioned above with the first group of parameters mentioned above. BENEFIT OF THE INVENTION
[0011] Making use of the time series standard of meteorological data in each particular growth phase, the invention makes it possible to carry out a high-precision harvest estimate, regardless of the existence of damage. SIMPLE EXPLANATION OF THE FIGURES
[0012] [Figure 1] It is the block diagram that shows the basic structure of the harvest forecasting system as the first form of implementation of this invention.
[0013] [Figure 2] It is the sequence diagram that shows the process performed with the harvest forecasting device in the first form of implementation of this invention.
[0014] [Figure 3] It is the flowchart that shows the process performed in the parameter priority calculation unit in the first form of implementation of this invention.
[0015] [Figure 4] It is the flowchart that shows the process performed in the image analysis unit in the first and second way of implementing this invention.
[0016] [Figure 5] It is the flowchart that shows the process performed in the unit of selection of agricultural land measured in the first and second form of implementation of this invention.
[0017] [Figure 6] It is the flowchart that shows the process performed in the harvest estimation unit in the first and second way of implementing this invention.
[0018] [Figure 7] It is an example of the GIS data structure for agricultural land in the first and second way of implementing this invention.
[0019] [Figure 8] It is an example of the data structure to the characteristic data of agricultural land, data of measured agricultural land and data of agricultural land not measured in the first and second form of implementation of this invention.
[0020] [Figure 9] It is the example of the data structure to the meteorological database in the first and second form of implementation of this invention.
[0021] [Figure 10] It is an example of the data structure to the data of the past in the first and second form of implementation, to the statistical data in the first form of implementation, and to the growth database in the first and second form of implementation of this invention.
[0022] [Figure 11] It is the explanatory diagram in the agricultural land selection unit measured in the first and second ways of implementing this invention.
[0023] [Figure 12] It is the explanatory diagram in the pixel extraction section inside agricultural land in those first and second forms of implementation of this invention.
[0024] [Figure 13] It is the explanatory diagram in the section of analysis of the time series patterns in the first and second forms of implementation of this invention.
[0025] [Figure 14] It is the input screen of the measured data entry unit in the first and second form of implementation of this invention.
[0026] [Figure 15] It is the display screen of the display unit for the result of the harvest estimate in the first and second form of implementation of this invention.
[0027] [Figure 16] It is the harvest forecasting system in the second form of implementation of this invention.
[0028] [Figure 17] It is the flowchart that shows the process performed in the growth phase classification section in that second form of implementation of this invention.
[0029] [Figure 18] It is an example of data structure to the database of the growth phase on agricultural land in that second form of implementation of this invention.
[0030] [Figure 19] It is the block diagram that shows the first and second ways of implementing the invention and the hardware of the harvest forecasting device.
[0031] [Figure 20] It is the sequence diagram that shows the first and second ways of implementing the invention and the flow of operations of this invention. WAY TO IMPLEMENT THE INVENTION
[0032] Explain a way of implementing this invention following the attached figures below.
[0033] This invention relates to the system for estimating the harvest of agricultural products on agricultural land. The agricultural land mentioned here indicates the agricultural areas within the predetermined area, including the spatially continuous agricultural areas that have the information of the same attribute.
[0034] In addition to harvesting, this invention can be applied when estimating the value of soil composition such as nitrogen, phosphoric acid, potassium etc., and the quantitative parameter such as the length of plants of agricultural products.
[0035] The explanation of the details related to this invention, in the example of the form of implementation below, this invention includes the contents below as an example.
[0036] Determine the group of parameters that apply as the evaluation criteria to select the measured agricultural land, through the previously accumulated images of the surveyed areas, the agricultural land attribute information, and the aerial images of the agricultural land such as research subject, then, select the measured agricultural land so that the mentioned parameter group has as many dispersions as possible, and also select those candidates from the measured agricultural land are concentrated in their most possible location, to reduce the probable research loads. In addition, analyzing the time series patterns of the meteorological data in each growth phase, calculating the parameter group related to the growth condition and performing the harvest estimate with the quantity of image characteristic, the attribute information of the agricultural land and the parameter group mentioned above as the explanatory variables.
[0037] This invention determines the parameter group that apply as the evaluation criteria to select the measured agricultural land, through the previously accumulated images of the surveyed areas, the agricultural land attribute information, and the aerial images of the agricultural land surveyed, then selects the measured agricultural land so that the mentioned parameter group has as many dispersions as possible, and also selects that the candidates of the measured agricultural land are concentrated in their most possible location, to reduce the probable research loads.
[0038] Agricultural land that appears to have the same spectrum in aerial images differs from models for estimating harvest depending on the difference in crop varieties, cultivation day, damage, and meteorological data. Therefore, this invention analyzes the time series patterns of the meteorological data in each growth phase, calculates the parameter group related to the growth condition and estimates the harvest with the amount of image characteristic, the attribute information of the terrains agricultural groups and the parameter group mentioned above as the explanatory variables.
[0039] In this invention, the inspector searches only the crop of agricultural land designated by the system, on the rest of the agricultural land, the estimate is made possible using the aerial images of the surveyed areas obtained by the sensor mounted on the satellite or on the aircraft, making it possible to reduce the inspector's load and the cost of research, and, improve the accuracy of harvest estimates.
[0040] Explain in detail the form of implementation of this invention relating the figures below. First form of implementation
[0041] Figure 1 is the block diagram showing the basic structure of the harvest forecasting system.
[0042] The harvest forecasting system in this form of implementation, the harvest forecasting device 10 is equipped.
[0043] The harvest forecasting device 10 equips the functions such as: after receiving the data before the last fiscal year obtained by the data from the past 11, the images of the agricultural land as the subject of the harvest estimate in the corresponding fiscal year through the image database 12, and the agricultural land data as the subject of the harvest estimate in the corresponding fiscal year by the GIS for agricultural land 13, transmit the information on the selected agricultural land as the measured subject to the local inspector 15, and; after receiving the result of the local survey by the local inspector 15, receiving the meteorological data from the time series in the corresponding fiscal year by the meteorological database 14 and outputting the result of the harvest estimate 17.
[0044] To perform the functions mentioned above, the harvest forecasting device 10, equips the unit for calculating the parameter priority 101, the image analysis unit 102, the unit for selecting the measured agricultural land 103, the farmland characteristic data 105, the measured farmland display unit 106, the measured data input unit 107, the measured farmland data 108, the unmeasured farmland data 109, the harvest estimate unit 110 and the display unit for harvest estimate result 111.
[0045] The unit of calculation of parameter priority 101, equips the function of receiving the data obtained before the last fiscal year by the data of the past 11, and transmitting the parameter priority used to select the agricultural land measured to the unit of selection of measured agricultural land 103. To perform the functions mentioned above, the unit for calculating the parameter priority 101 equips the statistical analysis section 1011, the priority calculation section 1012 and the statistical data 1013. Explain later the contents of concrete processes of each party and the data stored in the statistical data 1013.
[0046] Image analysis unit 102 receives the images of agricultural land as the subject of the harvest estimate in the corresponding fiscal year through the image database 12, and the data of agricultural land as the subject of the harvest estimate in the year corresponding tax by the GIS for agricultural land 13, and stores the amount of agricultural land spectrum characteristic calculated by the images in the agricultural land characteristic data 105. To perform the functions mentioned above, the image analysis unit 102 is equipped with pixel extraction section on agricultural land 1021 and the characteristic calculation section on agricultural land 1022. Explain later the contents of concrete processes of each part.
[0047] The measured farmland selection unit 103 receives the amount of agricultural land spectrum characteristic calculated by the images in the farmland characteristic data 105 and the parameter priority used as the decision material to select the measured farmland by the unit of calculation of the priority of parameter 101, and transmits, to the display unit of the measured agricultural land 106, the information of the measured agricultural land selected within the agricultural land existing in the surveyed areas. To perform the functions mentioned above, the measured farmland selection unit 103 equips the farmland selection axis determination section 1031, the histogram generation section 1032, the measurement candidate selection section 1033, the measurement cost calculation section 1034 and the determination section of the measured agricultural land 1035. Explain later the contents of the concrete processes of each party.
[0048] The measured farmland display unit 106 shows the inspector at site 15, the measured farmland information selected within the farmland existing in the surveyed areas received by the measured farmland selection unit 103. Explain the example later of the concrete display screen.
[0049] The measured data entry unit 107 receives the result measured by the inspector at site 15, the quantity of agricultural land spectrum characteristic by the agricultural land characteristic data 105 and updates the data stored in the measured agricultural land data 108 and data on unmeasured agricultural land 109.
[0050] The harvest estimate unit 110 receives data from agricultural land to the harvest estimate as the subject of the survey by the GIS for agricultural land 13, the time series patterns of the meteorological data in the corresponding fiscal year by the meteorological database 14 , receives the amount of spectrum characteristic of agricultural land and the yields obtained from the actual measurement by the measured farmland data 108 and the unmeasured farmland data 109, updates the data stored in the measured farmland data 108 and the data of unmeasured agricultural land 109, and the harvesting display unit for the result of the harvest estimate, transmits the ending flag (signal of completion) of the harvest estimate. To perform the functions mentioned above, the harvest estimation unit 110 is equipped with the growth phase estimation section 1101, the time series analysis section 1102, the model development section 1103, the estimation section of harvest 1104 and the growth database 1105. Explain later the contents of concrete processes of each party.
[0051] The harvesting display unit for harvest estimate result 111, after receiving the ending flagda harvest estimate by the harvest estimate unit 110, receives the yields obtained from the actual measurement by the measured farmland data 108, the yields obtained by the estimate using data from unmeasured agricultural land 109 and the result of the harvest estimate 17.
[0052] Explain later the data stored in the data of the past 11, in the GIS for agricultural land 13, in the meteorological database 14, in the characteristic data of agricultural land 105, in the data of measured agricultural land 108 and in the data of agricultural land unmeasured 109. In addition, the harvest forecasting device can equip data from the past 11, the image database 12, the GIS for agricultural land 13, and the weather database 14 inside.
[0053] The image database 12 stores the images obtained by the sensor maintained on the satellite or on the aircraft, including at least the filming date and the location information of the image in the header information. The shooting date indicates when that image data was obtained, and the location information for the image shows location information in the four directions. Additionally, the information about the location of the image, for example, can be the information with the latitudes and longitude in geodetic data. Perform the harvest estimate using mainly the aerial images covered in the image database 12.
[0054] Figure 2 is the sequence diagram showing an example of the process performed with the harvest forecasting device 10. Explain the concrete processes to follow.
[0055] First, the GIS for agricultural land 13 transmits the identification of agricultural land as the subject of the harvest estimate and the coordination information in the surveyed areas to the image analysis unit 102 (S201). At the same time, the image analysis unit 102 receives the images of agricultural land as the subject of the harvest estimate in the corresponding fiscal year through the image database 12 and calculates the amount of image characteristic of the respective agricultural land as the subject harvest estimate. Then, it stores the quantity of agricultural land image characteristics as the subject of the harvest estimate in the agricultural land characteristic data 105 (S202). The characteristic data of agricultural land 105, when storing the quantity of image characteristic of said agricultural land as the subject of the harvest estimate by the image analysis unit 102, transmits the quantity of image characteristic mentioned above to the selection unit of measured agricultural land 103 (S204). The parameter priority 101 calculation unit receives data obtained from data from the past 11 before the last fiscal year, to which data mentioned above performs the statistical analysis and calculates the parameter priority used as the decision material for selecting agricultural land measured, transmits the measured agricultural land selection unit 103 (S203). As for the sequential process of S201, S202 to S204 and the process of S203, it can treat any one simultaneously or preferably. The agricultural land selection unit measured03, after receiving the agricultural land information as the subject of harvest estimate by the GIS for agricultural land 13 (S205), the quantity of image characteristic mentioned above by the agricultural land characteristic data 105 and the parameter priority mentioned above by the parameter priority 101 calculation unit, selects the measured agricultural land, and transmits the selected information about the measured agricultural land to the local inspector 15 (S206). The local inspector 15 performs the actual measurement on the spot based on the information received in relation to the agricultural land measured to obtain the harvest on the respective agricultural land (S207). After obtaining the harvest on the respective farmland by the local inspector 15, the measured farmland data 108 and the unmeasured farmland data 109 receive the identification of the respective farmland and its image feature quantity by the land feature data agricultural 105 (S209) and, in addition, receive the harvest on the respective agricultural land by the local inspector 15 through the measured data entry unit 107 (S208). At the same time, in the measured agricultural land data 108, the quantity of characteristic of the image and the harvest corresponding to the identification of the agricultural land surveyed by the local inspector 15 are stored and, on the other hand, in the data of unmeasured agricultural land 109, store the amount of image characteristic and the harvest corresponding to the identification of the agricultural land not surveyed by the local inspector 15. The harvest stored in the data of unmeasured agricultural land 109, stores NULL due to the lack of information in the step of S208. After that, the harvest estimate unit 110 receives the identification, the quantity of the image characteristic and the harvest on the respective agricultural land by the measured farmland data 108 and by the unmeasured farmland data 109 (S210), receives the data of agricultural land as the subject of the harvest estimate by the GIS for agricultural land 13 (S211), and receives the time series standards of the meteorological data in the last fiscal year by the meteorological database 14 (S212) and develops the model for estimating the harvest. After that, the harvest on the respective agricultural land estimated based on the developed harvest estimate model is stored in the measured agricultural land data 108 and in the measured agricultural land data 109 (S213). After that, the harvest estimate unit 110 transmits the ending flag of ending harvest estimate to harvest display unit for harvest estimate result 111 (S215), the display unit for harvest estimate result 111 receives the identification, the harvest and estimated harvest of respective agricultural land by measured farmland data 108 and unmeasured farmland data 109 (S214). In the end, when harvesting the display unit for the result of the harvest estimate 111, it is shown in the result, which ends all processes in the harvest forecasting device 10.
[0056] Figure 3 is the flowchart that shows an example of the process performed in the parameter priority calculation unit. The parameter priority calculation unit receives the data before the last fiscal year obtained by the data of the past lie and transmits the parameter priority used as the decision material to select the measured agricultural land to the measured agricultural land selection unit 103. Explain the concrete processes to follow.
[0057] S31 shows the process start of the parameter 101 priority calculation unit.
[0058] In S32, section to analyze the 1011 statistic analyzes the data before the last fiscal year obtained by the data of the past lie and stores the 1013 statistical data. Specifically, on each parameter of those data obtained before the last fiscal year (1103 - 1111) , calculate the referred statistical values for each fiscal year. The statistical values mentioned here are indicated as the mean, the dispersion, the sample numbers, the histogram, the maximum value and the minimum value. The histogram mentioned here shows the degree corresponding to the class of the respective parameter. Additionally, the way of deciding the class can be arbitrary, for example, it can be decided by dividing the zone between the minimum and the maximum value in ten. If the aforementioned parameter is the discrete value of damage or varieties of crops, the average, dispersion, maximum and minimum values cannot be calculated and, therefore, NULL is stored.
[0059] In S33, the section for calculating the priority 1012 receives the annual statistical values of each parameter by the statistical data 1013, together with the name of each parameter and the value of the calculated priority to the unit of selection of the measured agricultural land 103. Specifically , compare the annual distribution of each parameter, when there are so many similar distributions, the priority is calculated at the high level, on the other hand, when the annual distributions are very different, the priority is calculated at the low level. For example, when the parameter is the continuous value, designate the total years as "I", the annual index as "i", the parameter used as "Mi", the distribution integrating all fiscal years as "V", and the sample numbers as “Ni”, the priority may be (Formula 1) below. [Formula 1]
[0060] Otherwise, it can be formulated with the distance of the distributions. In this case, if the parameter is the continuous value or the discrete value, the distance of the distributions can apply. For example, when the probability distribution of the fiscal year “i” parameter is designated as pi (x) and the probability distribution of the “x” parameter integrating all fiscal years as p (x), it can be (Formula 2) below. [Formula 2]
[0061] The symbols d (pi (x), pj (x)) here mean the measure that measures the distance between pi (x) and pj (x) of the probability distribution, for example, the divergence of Kullback-Leibler.
[0062] S34 shows the end of the process for calculating the priority of parameter 101.
[0063] Figure 4 is the flowchart showing an example of the process performed in the measured data entry unit 102. The measured data entry unit 102 receives images of agricultural land as the subject of the harvest estimate in the corresponding fiscal year. by the image database 12, and the farmland data as the subject of the harvest estimate in the corresponding fiscal year by the GIS for farmland 13, and stores the amount of farmland characteristic calculated by the images in the land characteristic data 105. Explain the specific processes to be followed.
[0064] S41 shows the process start of the measured data entry unit 102.
[0065] In S42, from the aerial images received by the image database 12 and the information on the form of the respective agricultural land, the pixel value of the aerial images included in the agricultural land is extracted. Explain the concrete method of extraction below.
[0066] In S43, if the number of pixels extracted during the S42 process time is more than 1, S44 advances, otherwise S45 advances.
[0067] In S44, the characteristic calculation section on agricultural land 1022 calculates the quantity of characteristics from the set of extracted pixel value, stores the identification of agricultural land 1051 corresponding to the characteristic data of agricultural land, the number of pixels interior of agricultural land in the amount of 1052 pixels and the amount of characteristic calculated interior of agricultural land in the amount of characteristic 1053. The number of pixels stored in the value of 1052 pixels can be an alignment of all the interior pixel values of agricultural land or use the average pixel value inside agricultural land. The same process applies to the amount of characteristic053. The quantity of characteristic mentioned here is the parameter obtained by implementing the predetermined operation at the pixel value, for example, you can use the Index of Normalized Difference Vegetation which is the ratio of the value obtained by subtracting the pixel values from the red of the pixel value near infrared and the sum of the number of near infrared pixels and the number of red pixels as the number of characteristics, or you can also use these parameters, segmenting the set of pixel values included inside agricultural land, and aligning the parameter calculated by the implementation of the predetermined operation to the average value of pixels belonging to the respective group and the parameter divided the number of pixels belonging to the respective group with the number of pixels inside the agricultural land, such as the number of characteristics.
[0068] In S45, due to the number of extracted pixels being 0, it stores the identification of agricultural land 1051 to the characteristic data of that agricultural land 105, storing NULL in the value of pixels 1052 and in the quantity of characteristics 1053.
[0069] In S46, make the decision to finalize. On all agricultural land including in the GIS for agricultural land 13, when making S44 or S45, advance to S47, if there is agricultural land never done both S44 and S45, advance to S42, extract the interior pixel value on agricultural land also on the nearby agricultural land.
[0070] S47 shows the end of the process of the measured data entry unit 102.
[0071] Figure 5 is the flowchart showing an example of the process performed in the measured farmland selection unit 103. The measured farmland selection unit 103 receives the amount of spectrum characteristics of agricultural land by the characteristic data of the agricultural land 105 and the parameter priority used as the decision material for selecting the agricultural land measured by the unit of calculation of the parameter priority 101, and transmits the information on the measured agricultural land selected in the agricultural land existing in the surveyed area to the unit of visualization of measured agricultural land 106. Explain the concrete processes to be followed.
[0072] S51 shows the start of the process for selecting the measured agricultural land 103.
[0073] In S52, receive that data stored by said data of characteristic of agricultural land 105 and by the GIS for agricultural land 13, and calculate the statistical values of the last fiscal year to the respective parameters. The aforementioned statistical values are the same data stored in the 1013 statistical data.
[0074] In S53, select some important parameters to the actual measurement. You can select the parameter whose priority received by the parameter priority 101 calculation unit is greater than the determined threshold, said parameter whose dispersion is greater than the threshold determined in the statistical values of the respective parameters calculated in S52, and moreover, on the respective parameters, calculate the value by multiplying the aforementioned priority with the dispersion value, on the conducted value, you can select the parameter that is greater than the determined threshold. In addition, if the parameter that has the statistical values calculated in S52 is not the continuous value, but the discrete value, it can use entropy instead of dispersion.
[0075] The processes of S52 and S53 are carried out by the section for determining the axis of selection of agricultural land 1031.
[0076] In S54, generate the multivariate histogram with axis of the parameters selected in S53.
[0077] The S54 process is performed by the histogram generation section 1032.
[0078] In S55, divide the axes of the multivariate histogram generated in S53 to the zones. On this occasion, the method of division or the number of division are arbitrary, for example, you can divide the zone of the respective parameters between the minimum value and the maximum value in ten, or you can divide the zone from the value subtracted from the constant factor between the mean, the median and the standard deviation up to the value added in the constant factor between the mean and the standard deviation in ten.
[0079] In S56, select the candidate of the measured agricultural land referring to the multivariate histogram divided the zones on the respective axes. Explain later the contents of concrete processes.
[0080] The processes of S55 and S56 are carried out by the selection section of measurement candidates 1033.
[0081] In S57, calculate the measurement costs for all candidate combinations of the measured agricultural land selected in S56. Measurement costs can be decided by the fact which of the distributions have the selected agricultural land spatially, notably, the shorter the distance between the mutual selected measured agricultural land is, the lower the measurement costs become, on the contrary, as the wider the aforementioned distance is, the higher the measurement costs become. It can be the determinant of the covariance matrix calculated by the set of coordinate information of the selected measured agricultural land (the average value of the coordinate data 1302 to the respective agricultural land) as the measurement costs.
[0082] The S57 process is carried out by the selection section of measurement candidates 1034.
[0083] In S58, select the candidate of the measured agricultural land that has less measurement costs calculated as the final measured agricultural land, transmitting to the display unit of the measured agricultural land 106. On this occasion, within the data stored in the GIS for agricultural land 13, transmits the list of agricultural land related to the identification of the selected agricultural land.
[0084] The S58 process is carried out by the section for determining the measured agricultural land 1035.
[0085] S59 shows the end of the process of selecting the measured agricultural land 103.
[0086] Figure 6 is the flowchart that shows an example of the process performed in the harvest estimate unit 110. Receive data from agricultural land as the subject of the harvest estimate by the GIS for agricultural land 13, the time series patterns of meteorological data in the last fiscal year by the meteorological database 14, the amount of spectrum characteristics of agricultural land and the harvest obtained in the actual measurement by measured farmland data 108 and unmeasured farmland data 109, updates the stored data in the data of unmeasured agricultural land 109, and transmits the ending flag of finishing the harvest estimate to the display unit for harvest estimate result 111. Explain the concrete processes to follow.
[0087] S61 shows the start of the harvest estimate unit process 110.
[0088] In S62, receive the meteorological data from the time series by the meteorological database 14 and estimate the time of the growth phase for the respective agricultural land. Specifically, to the respective agricultural land, on every day from the date of planting to the date of photographing the images or the date of harvest, calculate the accumulated temperature, the accumulated amount of solar radiation, as well as the accumulated amount of precipitation, compare with those data stored in the growth database 1105. To do this, on every day from the date of planting to the date of photographing the images or date of harvest, estimate which date is equivalent to the appropriate growth stage. On this occasion, the accumulated temperature, the accumulated amount of solar radiation and also the accumulated amount of precipitation apply as explanatory variables, and the growth phase applies as the response variables. Clustering or agglomeration of nearest k-neighbors, k-means or contaminated normal distribution can be applied as the estimation method. In addition, on this occasion, the meteorological data can be used with those meteorological data observed in the position closest to the corresponding agricultural land, or be interpolated with meteorological data of the position of corresponding agricultural land using the meteorological data of more than one place.
[0089] The S62 process is carried out by the growth phase estimation section 1101.
[0090] In S63, calculate the explanatory variables used to estimate the harvest using the meteorological data from the time series of the corresponding agricultural land. Calculate the explanatory variables by analyzing the time series patterns of the respective parameters of the meteorological data in each growth phase estimated in S61. Explain the details later.
[0091] The S63 process is performed by the pattern analysis section of the 1102 time series.
[0092] In S64, develop the harvest estimate models using the quantity of the spectrum characteristic and the harvest of the respective agricultural land received by the measured agricultural land data 108, and the explanatory variables based on the time series meteorological data calculated in S63 . On this occasion, the explanatory variables are based on the amount of characteristic of the spectrum of the respective agricultural land received by the measured agricultural land data 108 and on the time series meteorological data calculated in S63, the response variables are the harvest. In addition, there are several techniques used to develop the harvest estimation models. For example, when the index of explanatory variables like “i”, explanatory variables like “Xi”, and the response variables like “Y”, can generate the respective parameters by making the linear model like that of (Formula 3) or, taking the nonlinear models as those of (Formula 4) and (Formula 5).
[0093] Furthermore, you can use mixed models for crop varieties, or also use Bayesian Hierarchical Models.
[0094] The S64 process is carried out by the model development section 1103.
[0095] S65 assigns the amount of spectrum characteristic of the respective agricultural land received by the data of unmeasured agricultural land 109 and the explanatory variables based on the meteorological data of the time series calculated in S63, to the harvest estimate model developed in S64, and estimates the yields of the respective agricultural land. It then stores the estimated yields on the agricultural land corresponding to the estimated 1085 harvest of measured farmland data 108 and unmeasured farmland data 109. After that, transmits the ending flag to the display unit for harvest estimate result 111.
[0096] The S65 process is carried out by the harvest estimate section 1104.
[0097] S66 shows the end of the harvest estimate unit 110 process.
[0098] Figure 7 is an example of the GIS data structure for agricultural land 13. The GIS for agricultural land 13 stores data from 1301-1309, related to the coordinates, shape, and attribute information that are prepared in advance. .
[0099] The identification of agricultural land 1301 is the label to identify agricultural land as the subject of the harvest estimate.
[00100] The coordinate data 1302 is that which will be aligned with the coordinates at the respective vertices when the forms of agricultural land are considered as the polygon. The coordinates stored in coordinate data 1302 can be the arbitrary coordinate system, for example, it can store latitudes and longitudes in geodetic data.
[00101] Fertilizer 1303 shows the amount of fertilizer applied to the respective agricultural land. Fertilizer 1303 can be, for example, the amount of nitrogen, the amount of phosphoric acid and the amount of potassium applied before cultivation. The unit is arbitrary, for example, it can use kilograms for 10 ares.
[00102] The varieties of the 1304 crops are stored the names of the varieties of the crops of the agricultural products grown on the respective agricultural land. The varieties of crops 1304 may allow the names of the agricultural products stored together.
[00103] Damage 1305 stores the names caused on the respective agricultural land. If there is more than one damage to the single agricultural land, you can store the damage that has the greatest influence within the damage caused, or you can store all the damage caused.
[00104] Standard unit harvest 1306 stores the estimated crops that are harvested on average on the respective agricultural land. The unit is arbitrary, for example, it can use kilograms for 10 ares.
[00105] Planting phase 1307 stores the planting date of the corresponding fiscal year on the respective agricultural land.
[00106] The date of earing leveling 1308 stores the date of earing leveling for the corresponding fiscal year on the respective agricultural land.
[00107] The harvest date 1309 stores the harvest date of the corresponding fiscal year on the respective agricultural land.
[00108] Figure 8 is an example of the structure of the following data: farmland characteristic data 105, measured farmland data 108 and unmeasured farmland data 109. Farmland characteristic data 105 stores the 1051 - 1053 data related to the quantity of the image characteristic on the respective agricultural land. The measured farmland data 108 and the unmeasured farmland data 109 store the 1081 - 1085 data related to the image characteristic quantities and the harvest on the respective agricultural land.
[00109] The identification of agricultural land 1051 is the label to identify agricultural land as the subject of the harvest estimate, which agrees with the identification of agricultural land 1301 in the GIS for agricultural land 13.
[00110] Pixel value 1052 is the pixel value included in the corresponding agricultural land within the aerial images of the surveyed area. You can store all the pixel values included on agricultural land, or store the average pixel value on agricultural land.
[00111] The number of characteristics 1053 that is aligned with the parameters calculated by the predetermined operation from the 1052 pixel value. For example, one can make use of the Normalized Difference Vegetation index, which is the reason for the value obtained by subtracting the pixel values of the red of the near infrared pixel value and the sum of the number of near infrared pixels and the number of red pixels as the characteristic quantity, or you can also use these parameters, segmenting the set of pixel values including the interior of agricultural land, and aligning the parameter calculated by implementing the predetermined operation to the average value of pixels that belongs to each group and the parameter divided the number of pixels that belongs to each group with the number of interior pixels of agricultural land.
[00112] The identification of agricultural land 1081 is the label to identify agricultural land as the subject of the harvest estimate, which agrees with the identification of agricultural land 1301 in the GIS for agricultural land 13.
[00113] 1082 pixel value is the pixel value included in the corresponding agricultural land within the aerial images of the surveyed area, which agrees with the 1052 pixel value.
[00114] The number of characteristics 1083 that is aligned with the parameters calculated by the predetermined operation from the pixel value 1052, which agrees with the quantity of characteristic 1053.
[00115] Harvest 1084 is the aforementioned harvest of the corresponding agricultural land. The unit is arbitrary, for example, it can use kilograms for 10 ares. Agricultural land stored in unmeasured agricultural land data 109 is agricultural land. They are not really measured by the inspector, therefore, harvest 1084 of data from unmeasured agricultural land 109 is stored NULL.
[00116] The estimated harvest 1085 is the estimated harvest of the corresponding agricultural land. The unit is arbitrary, for example, it can use kilograms for 10 ares. The estimated harvest 1085 stores the estimated harvest by the harvest estimate unit 110. Therefore, immediately after the completion of S209, NULL is stored, and immediately after the completion of S213, the estimated harvest is stored.
[00117] Figure 9 is an example of the data structure for the meteorological database 14.
[00118] The meteorological database 14 stores data 1400 - 1412 related to the meteorological data of time series observed in the researched area. The meteorological database 14 stores the observed data in more than one position.
[00119] Observation identification 1400 is the identification to identify the observation position.
[00120] The observation point 1401 is the coordination information related to the corresponding observation position places. The coordination stored at observation point 1401 can be the arbitrary coordinate system, for example, it can store latitudes and longitudes in geodetic data.
[00121] Date 1402 shows the date when the meteorological data are observed.
[00122] The average temperature 1403 shows the average temperature at the observation point within the corresponding date. The unit is arbitrary, for example, it can employ degree Celsius (° C).
[00123] The maximum temperature 1404 shows the maximum temperature at the observation point within the corresponding date. The unit is arbitrary, for example, it can employ degree Celsius (° C).
[00124] Minimum temperature 1405 shows the minimum temperature at the observation point within the corresponding date. The unit is arbitrary, for example, it can employ degree Celsius (° C).
[00125] The amount of solar radiation 1406 shows the average amount of solar radiation at the observation point within the corresponding date. Additionally, just like the temperature, it can also store the maximum amount of solar radiation and the minimum amount of solar radiation. The unit is arbitrary, for example, it can employ watt per square meter (W / m2).
[00126] Humidity 1407 shows the average humidity at the observation point within the corresponding date. Additionally, just like the temperature, you can also store the maximum humidity and the minimum humidity. The humidity is arbitrary, for example, it can use percentage (%).
[00127] Atmospheric pressure 1408 shows the average atmospheric pressure at the observation point within the corresponding date. Additionally, even with temperature, it can also store the maximum atmospheric pressure and the minimum atmospheric pressure. The unit is arbitrary, for example, it can use percentage (%).
[00128] Precipitation 1409 shows the cumulative amount of precipitation at the observation point within the corresponding date. The unit is arbitrary, for example, it can use millimeter (mm).
[00129] The insolation duration 1410 shows the insolation duration at the observation point within the corresponding date. The unit is arbitrary, for example, it can use hour (h).
[00130] Wind orientation 1411 shows the wind orientation at the observation point within the corresponding date.
[00131] Wind speed 1412 shows the average wind speed at the observation point within the corresponding date. Additionally, it can also store the maximum instant wind speed. The unit is arbitrary, for example, it can use meter per second (m / s).
[00132] Additionally, the meteorological database 14 can have a table at each observation point like the examples mentioned above, or, adding observation identification 1400 and observation place 1402 to the column, it can store all data in the table singular.
[00133] Figure 10 is an example of the data structure for data from the past 11, the statistical data 1013, and the growth database 1105. The data from past 11 stores data 1101 - 1112, on the respective agricultural land. , the attribute information and the quantity of the image characteristic on the respective agricultural land that were obtained before the last fiscal year. Basically it agrees with the GIS for agricultural land 13 and the characteristic data of agricultural land 105, but it is only different in the point that the GIS for agricultural land 13 and the characteristic data of agricultural land 105 obtained in the corresponding fiscal year.
[00134] The identification of agricultural land 1101 is the label to identify agricultural land as the subject of the harvest estimate corresponding to the fiscal year, which agrees with the identification of agricultural land 1301.
[00135] The coordinate data 1102 are those that are aligned with the coordinates at the respective vertices, when the shapes of the agricultural land corresponding to the fiscal year are considered as the polygon, they agree with the coordinate data 1302.
[00136] Fertilizer 1103 shows the amount of fertilizer applied to the respective agricultural land corresponding to the fiscal year, according to fertilizer 1303.
[00137] The varieties of the 1104 crops agree with the varieties of the 1304 crops, storing the names of the varieties of the crops of the agricultural products grown on the respective agricultural land corresponding to the fiscal year.
[00138] Damage 1105 agrees with damage 1305, storing the names caused on the respective agricultural land corresponding to the fiscal year.
[00139] Standard unit harvest 1106 agrees with those standard unit harvests 1306, storing the estimated harvests that are harvested on average in the respective agricultural land corresponding to the fiscal year.
[00140] Planting phase 1107 agrees with planting phase 1307, storing the planting date on the respective agricultural land corresponding to the fiscal year.
[00141] The date of the leveling of the formation of the ear 1108 agrees with the date of the leveling of the formation of the ear 1308, storing the date of the leveling of the formation of the ear in the respective agricultural lands corresponding to the fiscal year.
[00142] Harvest date 1109 agrees with harvest date 1309, storing the harvest date on the respective agricultural land corresponding to the fiscal year.
[00143] The pixel value 1110, corresponding to the fiscal year, is the pixel value included in the corresponding agricultural land within the aerial images of the surveyed area, it is agreed with the pixel value 1052.
[00144] The number of characteristics 1111 is aligned with those parameters calculated by the predetermined operation from the 1110 pixel value, it is agreed with the 1053 pixel value.
[00145] The year 1112 shows the fiscal year in which the corresponding data are obtained.
[00146] Statistical data 1013 stores data from 10131 - 10138 of the respective years related to data from the past 11.
[00147] The year 10131 shows the year of obtaining the data in which the statistical value is calculated. It corresponds to the data 1112 stored in the data of the past 11.
[00148] Parameter 10132 shows the types of the parameter that calculates the statistical value. It corresponds to the column names of data 1102 - 1111 stored in the data of the past 11.
[00149] The average 10133 shows the average related to the corresponding parameter in the corresponding fiscal year. In addition, if the corresponding parameter is equivalent to the discrete value of the crop varieties or the damage, the average cannot be calculated, therefore, the NULL value is stored.
[00150] Dispersion 10134 shows the dispersion related to the corresponding parameter in the corresponding fiscal year. Likewise, if the corresponding parameter is equivalent to the discrete value of the crop varieties or the damage, the dispersion cannot be calculated, therefore, the NULL value is stored.
[00151] The number of samples 10135 shows the number of data related to the corresponding parameter in the corresponding fiscal year.
[00152] Area definition 10136 shows the class definition area in the histogram related to the corresponding parameter in the corresponding fiscal year. Explain specifically, if the parameter is the continuous value, the definition of area 10136 is equivalent to which the class numbers are aligned with the pairs of the maximum value and the minimum value of the definition area of the respective classes. If the parameter is the discrete value, the definition of area 10136 is equivalent to which all possible values are aligned.
[00153] Area A 10137 stores the data numbers existing in the first area definition within the data related to the corresponding parameter in the corresponding fiscal year.
[00154] Area A 10138 stores the data numbers existing in the second definition of area within the data related to the corresponding parameter in the corresponding fiscal year. Likewise, as area C, area D ..., there will be the same areas as the class number. For example, if the class number is 10, the column can exist up to area I. On the other hand, for example, if the discrete value is with only 5 types of the corresponding parameter, the NULL value is stored in the areas of F - 1.
[00155] The growth database 1105 stores the relationship of the respective growth phases with the accumulated temperature, the accumulated amount of solar radiation, the accumulated amount of precipitation and the varieties of crops as the database.
[00156] The growth phase 11051 shows the respective growth phases.
[00157] At the accumulated temperature 11052, the temperature accumulated at the time of the corresponding growth phase is stored. The unit is arbitrary, for example, it can employ degree Celsius (° C).
[00158] In the accumulated amount of solar radiation 11053, the accumulated amount of solar radiation is stored at the time of the corresponding growth phase. The unit is arbitrary, for example, it can employ watt per square meter (W / m2).
[00159] In the accumulated amount of precipitation 11054, the accumulated amount of precipitation is stored in the time of the corresponding growth phase. The unit is arbitrary, for example, it can use millimeter (mm).
[00160] The varieties of the 11055 crops show the information of the varieties of the crops in the referred data. In addition, you can add information about the types of agricultural products here.
[00161] Figure 11 is the explanatory drawing of the process carried out in the unit of selection of the measured agricultural land 103. Explain the concrete processes to follow.
[00162] In S56, divide the selected parameter axis and extract the data in a number decided by the respective generated grids. Figure 11 is the explanatory diagram that if the selected parameters are the standard unit harvest (axis 561) and the normalized vegetation index (axis 562).
[00163] First, on axes 561 and 562, generate the dispersion diagram 560, related to the data integrating the characteristic data of agricultural land 105 and the GIS for agricultural land 13. Then, divide axes 561 and 562 with the fixed number (562). At this time, the method of division or the number of division are arbitrary. For example, you can divide the parameter zone between the minimum and maximum value in ten, or you can divide the parameter zone from the value subtracted from the constant factor between the mean, the median and the standard deviation to the value added in the constant factor between the mean and the standard deviation in ten. Through the division, generating more than one 564 grid, extract the data in a fixed number by the respective grids (565). If the number of data in the grid is less than a fixed time, it will extract only the number of existing data. Then, select all extract combinations as candidates for the measured agricultural land, and advance to S57. For example, if the number of extracted data is 3, the total grid number is 4, and the data numbers in the respective grids are 10, 1, 2, and 5, there are 10! / (7! X 3!) Ways in the first grid, there are 1 way in the second and third grid. But, there are 5! / (2! X 3!) Ways in the fourth grid, with a total of 1200 ways of selection. These ways of selecting the data are the candidates of the measured agricultural land. On this occasion, the “! ”Means the factorial.
[00164] Figure 12 is the explanatory diagram of the processes in the section of extracting the pixels inside agricultural land 1021 (S42). Explain the concrete processes to follow.
[00165] Figure 12 is, in image 421, the example when extracting the pixels included in agricultural land 422. Pixel 423 is the set of pixels close to agricultural land 422.
[00166] When placing the information about the location of the four corners of the image 421, in the upper left: (sx 1, syl), in the upper right: (sx 2, syl), in the lower left: (sx 1, sy2), and in the lower right: (sx 2, sy2), and put the information on the location of the respective vertices of the agricultural land 422, (xl, yl), (x2, y2), (x3, y3), (x4, y4) .. ., comparing the position coordinates of the image 421 with the position coordinates of the agricultural land 422, it can be seen that the respective pixels within the images are included in the agricultural land or not. The pixels included in agricultural land are considered as the exit from the section of extracting the pixels inside agricultural land 1021. However, there are often pixels crossing agricultural land 422. In this case, just the case that the centers of each pixel is included in agricultural land, added to the above mentioned output.
[00167] Additionally, it depends on the position accuracy of the GIS for agricultural land 13 or the image database 12, this way of extraction can cause the images of the tracks outside the agricultural land to be mixed. Therefore, on agricultural land 422, you can extract the pixels inside the agricultural land, installing the buffer with the fixed distance 424 and making the new agricultural land 425.
[00168] Figure 13 is the explanatory diagram of the processes in the pattern analysis section of the 1102 (S63) time series. Explain the concrete processes to follow.
[00169] In S63, analysis of time series patterns of the respective meteorological data, calculation of parameters related to the growth of agricultural products and emission of those as explanatory variables used for the harvest estimate. Graph 6301 is the diagram that plots data such as date 6202 on the horizontal axis, and rainfall amount 6303 on the vertical axis. By estimating the growth phase at date 6302, you can divide the time series data into each growth phase of the planting date 6304. Additionally, the 6305 image capture date is the shooting date of the aerial images obtained by the images 12, is added to the data header information.
[00170] To analyze the time series patterns, first remove the high frequency components appearing as the noise of the graph 6301. Therefore, for example, you can solve the data using the moving average method and the low-pass filter. Graph 6306 is the time series standard obtained by eliminating high frequency components.
[00171] Issue the peak value 6307, the width of the respective peaks 6308 (for example, dispersion), the position of peaks 6309, and the cumulative value 6310 as the explanatory variables used to estimate the harvest. What's more, before and after the 6305 image capture date, you can calculate the aforementioned parameters separately and add those as the explanatory variables.
[00172] Explain these benefits by analyzing the time series patterns below.
[00173] Graph 6311 shows the time series changes of the growth parameters in the respective agricultural land in the graph. The parameter mentioned here, for example, is the parameter that has good growth and positive correlation, and for example, it can use the Index of Normalized Difference Vegetation calculated by the images.
[00174] Curve 6312 is the growth curve for agricultural land suffered from damage in the middle of the day despite the earlier cultivation day, and curve 6313 is the growth curve for undamaged agricultural land despite the day of cultivation. cultivation later. For example, When the images are obtained at the intersection point of curves 6312 and 6313, which are named 6314, it is difficult to estimate the harvest by the spectrum of the images. Therefore, it can perform to improve the accuracy of the harvest estimate, reproducing the growth curve before and after the date of the image capture by the meteorological data of the time series and obtaining the explanatory variables. For example, the sequence of rain from the time of the ear date to the time of the harvest date can cause a reduction in the harvest with the damage of fallen ears. On the other hand, the high temperature sequence at the time of the panicle differentiation phase can cause the greatest obstacle, and influences the growth curve. Therefore, use the parameters of 6307 - 6310 calculated by the time series standards, in each growth phase as the explanatory variables for the harvest estimate.
[00175] Additionally, as the meteorological data placed in the process of Figure 13, placing the data that estimates the future as the weather forecast, it is possible to forecast the harvest of the corresponding fiscal year before cultivation. Therefore, implementing the harvest estimate by simulating each condition in variety, for example, crop varieties and the day of cultivation, based on the result, can be applied to help agriculture.
[00176] Figure 14 is an example of the input screen of the measured data input unit 107. Explain the details of the input screen below.
[00177] Screen 10710 is the diagram showing the polygons of agricultural land. By showing the polygons, you can superimpose the polygons of agricultural land on top of aerial images like satellite images.
[00178] Agricultural land 10701 is the agricultural land now selected, and agricultural land 10702 is the agricultural land not selected. The information on agricultural land 10701 is shown in table 10703. What is shown in table 10703 is the data stored in the GIS for agricultural land 13 and the measured harvest of agricultural land 10704 surveyed by the local inspector. You can enter the measured harvest data 10704 via the keyboard.
[00179] Selecting the arrows 10705 - 10708, you can change the area of the map shown on screen 10710 in the direction of the arrow.
[00180] By selecting buttons 10710 - 10717, you can implement the respective processes. On this occasion, you can select the buttons by just touching the screen as the touchscreen, or you can select the buttons by the mouse cursor10709 and keyboard input.
[00181] Selecting button 10710, you can change the way the polygons shown on screen 10700 are displayed. For example, you can show by changing the color and line thickness of the polygons depending on the attribute value in table 10703, and more, on the measured harvest 10704, you can change the color and line thickness of the polygons closing between the agricultural land already entered and the non-entered. Also, on screen 10700, overlapping the identification of the corresponding agricultural land on top of the polygons, you can change the type of overlapping attribute.
[00182] By selecting the 10711 button, you can magnify the display scale of the 10700 screen. By selecting the 10712 button, you can reduce the display scale of the 10700 screen.
[00183] Selecting button 10713, you can search for the polygons of agricultural land for each attribute value. After searching, the result close to the agricultural land sought is shown on screen 10700.
[00184] With the button 10714, you can add the new data to the GIS for agricultural land 13, the data of measured agricultural land 108 and the data of unmeasured agricultural land 109. Adding to table 10703 via keyboard, the agricultural land is added.
[00185] With button 10715, you can update the data to the GIS for agricultural land 13, the data of measured agricultural land 108 and the data of unmeasured agricultural land 109. Adding to table 10703 via keyboard, the agricultural land is updated.
[00186] With button 10716, you can delete the data for the GIS for agricultural land 13, the data for measured agricultural land 108 and the data for unmeasured agricultural land 109. The data for the agricultural land now selected is deleted.
[00187] Button 10717 can cancel data entry. This operation ends the data entry by the local inspector.
[00188] Figure 15 is an example of the output screen of the display unit for the result of the harvest estimate. Explain the details of the output screen below.
[00189] The screen 11100, to the agricultural land 11101 and 11102, the table 11103, to the harvest measured 11104, the arrows 11105 - 11108 and the buttons 11110 - 11113 are respectively equivalent to the screen 10700, to the agricultural land 10701 and 10702, to the table 10703, the measured harvest 10704, the arrows 10705 - 10708 and the buttons 10710 - 10713.
[00190] Estimated harvest 11114 is the estimated harvest data 1085 stored in the measured farmland data 108 and the unmeasured farmland data 109.
[00191] The 11115 button implements the precision assessment. The agricultural land stored in the measured agricultural land data 108 is the only agricultural land having the measured harvest 11104 and the estimated harvest 11114, therefore, to the agricultural land, comparing the measured harvest with the estimated harvest, calculate the statistical value and implement the evaluation of precision. For example, you can calculate the MAE (Mean Average Error), that is, the average value of the change in the measured harvest and the estimated harvest, and you can calculate the proportion of agricultural land whose change in the measured harvest and the estimated harvest is within the threshold. fixed, and more, can calculate the correlation coefficient of the measured harvest and the estimated harvest.
[00192] Button 11116 generates the histogram related to the measured harvest 11104 and the estimated harvest 11114. The histogram mentioned here is the one that presents the degree dependent on the class of the corresponding parameter as the graph. Additionally, the way of deciding the class can be arbitrary, for example, it can be decided by dividing the zone between the minimum and the maximum value in ten.
[00193] Button 11117 leaves the attribute name and attribute value included in table 11103 as the CSV file (Comma-Separated Values'). On this occasion, by entering the conditions of the attribute value of the output data, you can filter the output data.
[00194] Figure 19 is the block diagram showing an example of the hardware structure of the harvest forecasting system.
[00195] The crop forecasting device 10 is the computer that has the 1901 operating unit, the 1902 display unit, the 1903 processor, the 1904 main memory and the 1905 storage device.
[00196] The 1903 processor performs the programs stored in the main memory 1904.
[00197] Main memory 1904, for example, whether it is a semiconductor memory, stores the programs carried out by the 1903 processor and the data referred to by the 1903 processor. Specifically, at least a part of the programs and data stored in the 1905 storage device, it is copied to main memory 1904, when necessary.
[00198] The 1903 processor, operating in accordance with the programs of the respective function sections, functions as the function section that performs the determined function. For example, working in accordance with the harvest forecasting program, the 1903 processor has its function as the 110 harvesting estimate unit. Other programs are also equivalent to that. The crop forecasting device 10 is the device that includes these function sections.
[00199] The 1901 operation unit receives the input operation by the user. The operating unit, for example, can include the keyboard or mouse.
[00200] The 1902 display unit leaves the information to the user. The 1902 display unit, for example, is the image display device like the liquid crystal display.
[00201] The 1905 storage device, for example, is the non-volatile storage device like the hard disk device (HDD) or flash memory. The storage device 1905 in this implemented form stores at least the parameter priority 101 calculation unit, the image analysis unit 102, the measured farmland selection unit 103 and the harvest estimate unit 110. Additionally, the respective databases and storage units can also be stored in the 1905 storage device.
[00202] And more, the information of the programs and databases that perform the respective functions of the object recognition device 10, can be stored in the storage devices like the storage device 1505, the non-volatile semiconductor memory, the hard disk device, the SSD (Solid State Drive), or, in non-temporary computer-readable data storage media such as the smart card, the Secure Digital Card DVD.
[00203] Figure 20 is the sequence diagram showing an example of the flow of operations of this invention. Explain the concrete flow below.
[00204] First, customer 304 makes the harvest estimate 305 order to the harvest estimate agency 303. On this occasion, it will provide the GIS data for agricultural land that customer 304 maintains. After that, the crop estimation agency 303 orders the photographing images 307 to the aerial photography photography agency 302. The aerial photography photography agency 302, upon receiving the request to photograph the images, does the photography instruction 308 to the server satellite 301. Then, you start shooting the 309 images. At the same time that the image is being photographed, that image is transmitted 310 to the aerial photography agency 302. Sequentially, the image is transmitted 310 to that estimation agency. crop 303, and also that image is transmitted 312 to client 304. After that, client 304 decides whether to perform the harvest estimate with the photographed image 313. For example, if the great cloudiness is considered in the image, decide not to use this image 314, on this occasion, the harvest estimation agency 303 orders again to photograph images 315. After that, repeat flow 308 - 303. After 312, if you decide to use image 31 6, the harvest estimation agency 303, through the harvest estimation device 10, performs the harvest estimate 317. Additionally, process 306 can be performed after process 316 and before process 317. After that, the result 318 is transmitted to client 304. Second form of implementation
[00205] Explain the second way of implementing this invention by referring to the figures below.
[00206] In the first form of implementation of this invention, selecting the agricultural land measured by the variations of data accumulated in the past, transmit the list of the agricultural land measured to the local inspector. However, there is a case without the accumulated data in the past, now, in the second form of implementation, select the agricultural land measured using only the data from the business year to perform the harvest estimate. Therefore, using aerial time series images and altitude data, estimate the growth phase at the same time as the aerial images used to estimate the harvest, use that growth phase as the parameter used when selecting the terrains agricultural
[00207] In the second form of implementation, the components are the same as those in the first form, except the unit for calculating parameter priority 101 and data from the past 11. Specifically, Figures 4-9ell-15 also apply in the second form. Explain only the differences to the first form below.
[00208] Figure 16 is the block diagram showing the basic structure of the harvest forecasting system.
[00209] The harvest forecasting system in this form is equipped with the harvest forecasting device 20.
[00210] The harvest forecasting device 20 is equipped with functions such as: the function that, after receiving the DEM data (Digital Elevation Model} of the areas as the subject by the altitude database 21, the aerial images of time series in areas such as the subject by the database of images in chronological series 22, the images of the corresponding fiscal year that the agricultural land as the subject of the harvest estimate are photographed by the image database 12, and the data of the agricultural land as the subject of the GIS harvest estimate for agricultural land 13, transmits the information on the selected agricultural land as the measured subject to the local inspector 15; the function that, after receiving the result of the local survey by the local inspector 15 and then after receiving the time series meteorological database for the corresponding fiscal year through the meteorological database 14, the result of the harvest estimate 17 comes out.
[00211] To perform the functions mentioned above, the harvest forecasting device 10 is equipped as follows: the growth phase classification section 201, the image analysis unit 102, the agricultural land selection unit measured 103, farmland characteristic data 105, measured farmland display unit 106, measured data input unit 107, measured farmland data 108, unmeasured farmland data 109, unity harvest estimate 110 and the display unit for harvest estimate result 111.
[00212] The following functions are the same as those in the first form: the growth phase classification section 201, the image analysis unit 102, the measured farmland selection unit 103, the farmland characteristic data 105, the measured farm display unit 106, the measured data entry unit 107, the measured farmland data 108, the unmeasured farmland data 109, the harvest estimate unit 110 and the display unit for result harvest estimate 111.
[00213] The growth phase classification section 201 equips the function that receives the DEM (Digital Elevation Model} data from the areas as the subject by the altitude database 21, and the aerial time series images in the areas as the subject by the chronological series image database 22, and then, transmits the parameter priority used in the selection of the measured agricultural land to the selection unit of the measured agricultural land 103. To perform the functions mentioned above, the classification section growth phase 201 equips the 2011 climatic variation estimation section, the accumulated temperature estimation section 2012, the growth phase estimation section 2013, and the database for the growth phase on agricultural land 2014. Explain later the contents of concrete processes.
[00214] DEM data from areas such as the subject are stored in the altitude database 21, and aerial chronological series images in areas as the subject in the chronological series image database 22. Whatever data are the data in the same way as that of the aerial images stored in the image database 12. By the way, the respective pixel values designate the altitude values in the appropriate position coordinates. The unit is arbitrary, for example, it can use meter (m).
[00215] Figure 17 is the flowchart that shows an example of the process performed in the growth phase classification section 201. Explain the contents of concrete processes to follow.
[00216] S20101 shows the departure of the process of classification section of growth phase 201.
[00217] In S20102, estimate the time series meteorological data, for example, the temperature, in the position of the respective agricultural land, by the DEM data stored in the altitude database 21, by the meteorological data stored in the meteorological database 14 and aerial time series images stored in the time series image database 22.
[00218] The place of observation of meteorological data stored in the meteorological database 14, is not always close to agricultural land, therefore, using the meteorological data appropriate to the place of observation closest to the corresponding agricultural land, causes a lot of difference in real weather data on agricultural land. Then, using DEM data and time series aerial images, estimate the meteorological data from the planting date to the harvest date at the position of the corresponding agricultural land. For example, through the pixel value of DEM data as “Dj” corresponding to the place of observation of meteorological data with order j as “Tj”, the date of observation of meteorological data with order j as “Tj”, and the value of pixels of aerial images A as “Aj” corresponding to the observation place, developing the linear regression equation as [Formula 6], calculating the, / and y, and applying the aforementioned formula to all agricultural land, can estimate time series meteorological data. [Formula 6]
[00219] For example, to solve the parameters of «, p and y, you can apply the least squares method.
[00220] The S20102 process is carried out by the 2011 climatic variation estimation section.
[00221] In S20103, calculate the accumulated value of the meteorological data from the date of cultivation or the date of ear formation of agricultural land until the date of photographing of the aerial images used for the harvest estimate.
[00222] The S20103 process is carried out using the accumulated temperature estimate section 2012.
[00223] In S20104, estimate the growth phase of the respective agricultural land using the accumulated value of meteorological data calculated in S20103. The estimation method can employ the method described in S62. After that, the identification of agricultural land and the estimated growth phase are stored in the database for the growth phase in agricultural land 2014.
[00224] In S20105, on the growth phase, calculate the priority. By the way, the parameter that exists is only the growth phase, observing the value that the growth phase may vary, it is enough to decide whether to use the value or not. For example, looking at the database for the growth phase on agricultural land 2014, if there is only one type of growth phase, which is considered to have no variation in the growth phase, determine that the priority is 0, and in others cases, the priority is 1.
[00225] Processes S20104 and S20105 are carried out using the 2013 growth phase estimation section.
[00226] S20106 shows the completion of the process of classification section of growth phase 201.
[00227] Figure 18 is an example of the data structure for the growth phase on agricultural land 2014.
[00228] The data for the phase of growth in agricultural land 2014 stores the result of estimation of the growth phase in the respective agricultural land, relating to the time when the aerial images capture for the harvest estimate.
[00229] The identification of agricultural land 20141 is the label to identify agricultural land as the subject of the harvest estimate, which agrees with the identification of agricultural land 1301.
[00230] The 20142 growth phase shows the corresponding growth phase.
[00231] The contents explained in the examples of the implementation method, limiting the number of agricultural land surveyed as much as possible, make it possible to estimate the high precision harvest, using aerial images and time series meteorological data in the special phase of product growth agricultural EXPLANATION OF SYMBOLS 10 Harvest forecasting device 11 Past data 12 Image database 13 GIS for agricultural land 14 Weather database 15 Local inspector 17 Harvest estimate result 21 Altitude database 22 Image database in chronological series 101 Parameter priority calculation unit 102 Image analysis unit 103 Measured farmland selection unit 105 Agricultural land characteristic data 106 Measured farmland display unit 107 Measured data entry unit 108 Measurement data measured agricultural land 109 unmeasured agricultural land data 110 harvest estimate unit 111 display unit for harvest estimate result 201 growth phase classification section 1011 statistical analysis section 1012 priority calculation section 1013 statistical data 1021 Pixel extraction section on agricultural land 1022 Section the calculation of characteristic on agricultural land 1031 Determination section of agricultural land selection axis 1032 Histogram generation section 1033 Measurement candidate selection section 1034 Measurement cost calculation section 1035 Determination of measured agricultural land section 1101 Growth phase estimation section 1102 Time series analysis section 1103 Model development section 1104 Crop estimate section 1105 Growth database 1901 Operating unit 1902 Display unit 1903 Processor 1904 Main memory 1905 Device storage 2011 Temperature variation estimation section 2012 Cumulative temperature estimation section 2013 Growth phase estimation section 2014 Database for growth phase on agricultural land

权利要求:
Claims (13)
[0001]
1. Harvest forecasting system, comprising: a receiving unit that receives aerial images including agricultural land, a storage unit that stores, respectively, form information including said agricultural land and information on the location of the agricultural land mentioned above, a image analysis unit (102) which calculates quantity of image characteristic of the mentioned agricultural land using the aerial images received that include the agricultural land mentioned above with the information of forms and the location information of said agricultural land, and a unit of image harvest estimate (110) which calculates expected harvest of the aforementioned agricultural land from that amount of calculated image characteristic, this system being characterized by the fact that the aforementioned harvest estimate unit (110) has an analysis section of the time series standards (1102) which, at each stage of particular growth of agricultural products grown on the aforementioned agricultural land, gives a first set of parameters related to the growth condition of the aforementioned agricultural products based on a time series pattern of previously stored meteorological data, and calculates an expected harvest of those agricultural land mentioned above through the quantity of image characteristic mentioned above with that first group of parameters mentioned above.
[0002]
2. Harvest forecasting system according to claim 1, characterized by the fact that the group of parameters is calculated by more than one of the following accumulated values: peak values, locations, numbers, and time series patterns of meteorological data , these accumulated values being extracted from the aforementioned time series standards excluding high frequency components.
[0003]
3. Harvest forecasting system according to claim 1, characterized by the fact that it additionally comprises: a parameter priority calculation unit (101) that calculates a parameter priority to select the agricultural land measured by the attribute information for each of the respective agricultural land stored in the aforementioned storage unit before the corresponding fiscal year, and a measurement farm selection unit (103) which selects the agricultural land measured by the aforementioned priority, quantity of image feature mentioned above and information on the aforementioned location of said agricultural land.
[0004]
4. Harvest forecasting system according to claim 1, characterized by the fact that it additionally comprises: a growth phase classification section (201) that estimates a growth phase of the crops mentioned above when acquiring the aerial images above above, in which the growth phase estimate mentioned above is calculated by time series meteorological data with time series aerial images or altitude data, the measured farmland selection unit (103) which selects the farmland measured by the following objects: estimation of the growth phase mentioned above, quantity of image characteristics mentioned above and information on the location of those agricultural lands mentioned above.
[0005]
5. Harvest forecasting system according to claim 3 or 4, characterized by the fact that it also comprises a measured data entry unit (107) that has the capacity to enter the measured harvest of the selected agricultural land measured above.
[0006]
6. Harvest forecasting system according to claim 1, characterized in that the section of analysis of the time series patterns (1102) mentioned above gives a second group of parameters related to the growth condition of the agricultural products mentioned above, at each stage of particular growth of agricultural products grown on such agricultural land mentioned above, before and after each date of photography of aerial images, and acquire time series patterns of meteorological data.
[0007]
7. Harvest forecasting system according to claim 3, characterized by the fact that the parameter priority calculation unit (101) calculates the aforementioned priority by comparing statistical information from the attribute information of agricultural land in each year before the corresponding fiscal year.
[0008]
8. Harvest forecasting system according to claim 1, characterized in that the aforementioned harvesting estimation unit (110) includes a growth phase estimation section (1101) that estimates a change date for the phase growth of agricultural land mentioned above from the time series meteorological data for the corresponding fiscal year.
[0009]
9. Harvest forecasting system according to claim 1, characterized by the fact that it also includes a display unit for the harvest estimate result (111) that shows the predicted harvest mentioned above.
[0010]
10. Crop forecasting device (10), comprising: an image analysis unit (102) that gives the quantity of agricultural land image characteristic from form information and information on the location of said agricultural land and aerial images photographed from the corresponding agricultural land, and a crop estimation unit (110) that gives an expected harvest of agricultural land based on the quantity of image characteristic, a device that is being characterized by further comprising a section for analyzing time series patterns (1102 ) which, at each stage of particular growth of agricultural products grown on the aforementioned agricultural land, gives a group of parameters related to the growth condition of the agricultural products mentioned above based on a time series pattern of previously stored meteorological data, and calculates expected harvest of those agricultural land mentioned above using the amount of image feature mentioned above and that parameter group mentioned above.
[0011]
11. Crop forecasting device (10) according to claim 10, characterized by the fact that it also includes a display unit for the result of the crop estimate (111) which shows the predicted harvest mentioned above.
[0012]
12. Crop forecasting device (10) according to claim 10, characterized by the fact that it also comprises a storage unit that stores the information of forms and the location information of the agricultural lands mentioned above.
[0013]
13. Crop prediction device (10) according to claim 10, characterized by the fact that it also comprises a reception unit that receives the aerial images mentioned above.

类似技术:

---

公开号 | 公开日 | 专利标题

---

BR112015031241B1|2020-07-07|crop forecasting system and device

---

French et al.2015|Remote sensing of evapotranspiration over cotton using the TSEB and METRIC energy balance models

---

Fu et al.2014|Estimating landscape net ecosystem exchange at high spatial–temporal resolution based on Landsat data, an improved upscaling model framework, and eddy covariance flux measurements

---

Möller et al.2017|Coupling of phenological information and simulated vegetation index time series: Limitations and potentials for the assessment and monitoring of soil erosion risk

---

Hou et al.2011|Use of ALS, Airborne CIR and ALOS AVNIR-2 data for estimating tropical forest attributes in Lao PDR

---

Huang et al.2020|The migration of training samples towards dynamic global land cover mapping

---

Gosset et al.2018|Evaluation of TAPEER daily estimates and other GPM‐era products against dense gauge networks in West Africa, analysing ground reference uncertainty

---

Jiang2013|GIS-based time series study of soil erosion risk using the Revised Universal Soil Loss Equation | model in a micro-catchment on Mount Elgon, Uganda

---

Franklin et al.2011|Landsat-comparable land cover maps using ASTER and SPOT images: a case study for large-area mapping programmes

---

Yang et al.2020|Remote sensing estimation of the soil erosion cover‐management factor for China's Loess Plateau

---

Fu et al.2021|Assessment of approaches for monitoring forest structure dynamics using bi-temporal digital aerial photogrammetry point clouds

---

Hollaus et al.2015|Site index assessment based on multi-temporal ALS data

---

Upgupta et al.2015|Estimation of aboveground phytomass of plantations using digital photogrammetry and high resolution remote sensing data

---

Aldhshan et al.2019|Change detection on land use/land cover and land surface temperature using spatiotemporal data of Landsat: a case study of Gaza Strip

---

Dai et al.2014|Estimation of daily winter precipitation in the Snowy Mountains of southeastern Australia

---

Chen et al.2010|Estimation of forest height, biomass and volume using support vector regression and segmentation from lidar transects and Quickbird imagery

---

Sarkar et al.2020|Assessment of agricultural drought in upper Seonath Sub-Basin of Chhattisgarh | using remote sensing and GIS-based indices

---

Belford et al.2014|Using remote sensing and geospatial analysis to understand changes to Lake Michigan dunes

---

Waser2012|Airborne remote sensing data for semi-automated extraction of tree area and classification of tree species

---

Nowak2013|A guide to assessing urban forests

---

Nakajima2016|Estimating tree growth using crown metrics derived from lidar data

---

KR102118802B1|2020-06-26|Method and system for mornitoring dry stream using unmanned aerial vehicle

---

Nakajima et al.2011|A growth prediction system for local stand volume derived from lidar data

---

Huang et al.2012|Mapping forest above-ground biomass and its changes from LVIS waveform data

---

Giffen et al.2014|Alaska: Glaciers of Kenai Fjords National Park and Katmai National Park and Preserve

同族专利:

---

公开号 | 公开日

---

JP2015000049A|2015-01-05|

---

BR112015031241A2|2017-07-25|

---

JP6147579B2|2017-06-14|

---

WO2014203664A1|2014-12-24|

---

BR112015031241B8|2020-07-28|

引用文献:

---

公开号 | 申请日 | 公开日 | 申请人 | 专利标题

---

---

JP2001256289A|2000-03-14|2001-09-21|Shigenao Daimatsu|Production managing system for agriculture and fisheries|

---

JP2003006274A|2001-06-25|2003-01-10|National Agricultural Research Organization|Judgement system for suitable crop and its kind to be introduced, and cultivation ground and time based on climate similarity at each growth stage|

---

JP4140052B2|2002-05-28|2008-08-27|株式会社サタケ|Image correction method for crop growth diagnosis method|

---

JP2005085059A|2003-09-10|2005-03-31|Sec:Kk|Prediction system for farmwork determination support|

---

JP4873545B2|2006-05-16|2012-02-08|株式会社日立ソリューションズ|Field management support method and system|

---

JP2010166851A|2009-01-22|2010-08-05|Chiharu Hongo|Method and device for predicting crop yield|JP6285875B2|2015-01-06|2018-02-28|株式会社日立製作所|Plant growth analysis system and plant growth analysis method|

---

US10028426B2|2015-04-17|2018-07-24|360 Yield Center, Llc|Agronomic systems, methods and apparatuses|

---

CN104914828A|2015-04-22|2015-09-16|柳州易农科技有限公司|Intelligent system for agriculture|

---

US10791666B2|2015-06-08|2020-10-06|The Climate Corporation|Agricultural data analysis|

---

US10312986B2|2015-08-19|2019-06-04|Intel Corporation|Flexible CSI RS configuration for FD-MIMO systems|

---

US11062223B2|2015-12-02|2021-07-13|The Climate Corporation|Forecasting field level crop yield during a growing season|

---

JP6639300B2|2016-03-24|2020-02-05|株式会社日立ソリューションズ東日本|Apparatus for estimating normal crop ratio of crops, apparatus for estimating crop yield of crops, and method for estimating normal crop rate of crops|

---

JP6679102B2|2016-03-25|2020-04-15|Ｎｅｃソリューションイノベータ株式会社|Shipment quantity forecasting device, shipment quantity forecasting method, and program|

---

JP6685856B2|2016-06-30|2020-04-22|株式会社日立製作所|Farm work plan support device and farm work plan support method|

---

WO2018151196A1|2017-02-16|2018-08-23|Ｎｅｃソリューションイノベータ株式会社|Crop growth estimation device, crop growth estimation method, program, and recording medium|

---

JP6333456B1|2017-08-30|2018-05-30|アビームコンサルティング株式会社|Production management apparatus, production management method and program|

---

US10891482B2|2018-07-10|2021-01-12|Adroit Robotics|Systems, devices, and methods for in-field diagnosis of growth stage and crop yield estimation in a plant area|

---

CA3125793A1|2019-02-05|2020-08-13|Farmers Edge Inc.|Harvest confirmation system and method|

---

CN113544714A|2019-03-08|2021-10-22|Assest 株式会社|Crop yield prediction program and cultivation environment determination program|

---

JP6704148B1|2019-03-08|2020-06-03|Ａｓｓｅｓｔ株式会社|Crop yield forecast program and crop quality forecast program|

---

WO2021157032A1|2020-02-06|2021-08-12|株式会社ナイルワークス|Growth assessment system, growth assessment server and growth assessment method|

---

CN112348229A|2020-10-12|2021-02-09|上海微亿智造科技有限公司|Product yield prediction method and prediction system based on industrial big data|

法律状态:
2018-02-27| B06F| Objections, documents and/or translations needed after an examination request according art. 34 industrial property law|

---

2019-07-09| B06T| Formal requirements before examination|

---

2020-05-05| B09A| Decision: intention to grant|

---

2020-07-07| B16A| Patent or certificate of addition of invention granted|Free format text: PRAZO DE VALIDADE: 20 (VINTE) ANOS CONTADOS A PARTIR DE 19/05/2014, OBSERVADAS AS CONDICOES LEGAIS. |

---

2020-07-28| B16C| Correction of notification of the grant|Free format text: REF. RPI 2583 DE 07/07/2020 QUANTO AO ENDERECO. |

优先权:

---

申请号 | 申请日 | 专利标题

---

JP2013127094A|JP6147579B2|2013-06-18|2013-06-18|Yield prediction system and yield prediction device|

---

JP2013-127094|2013-06-18|

---

PCT/JP2014/063150|WO2014203664A1|2013-06-18|2014-05-19|Harvest-predicting system and harvest-predicting device|

---