• 专利附录
  • 专利说明
  • 权利要求
  • 类似技术
  • 同族专利
  • 引用文献
  • 法律状态
  • 优先权
    • 专利摘要:
    A method (100) of classifying plants (24) of a field (22) comprising: - grasping the plants (24) of a segment (20) of the field (22) with the aid of an input unit optical and / or infrared (14), - determine (106) the analogy between the seized plants (24) in the field segment (20), with respect to each other and / or with respect to the plants (24 ' ) of another field segment (20 ') and / or a reference element using the plant image information, and - classifying (108) the plants entered (24) according to the degree of analogy .
    公开号:FR3071644A1
    申请号:FR1858775
    申请日:2018-09-26
    公开日:2019-03-29
    发明作者:Markus Hoeferlin
    申请人:Robert Bosch GmbH；
    IPC主号:
    专利说明:

    Field of the invention
    The present invention relates to a method for classifying plants growing on a field and to a classification device or unit applying this method. Finally, the invention relates to a control device and a computer program for implementing the method.
    State of the art
    To automate the various agricultural and seed production works such as the regulation of certain wild herbs or herbs, counting and measuring plants to characterize phenomena, selective harvesting, selective application of herbicides, fungicides, pesticides and insecticides and others, it is essential to classify plants, that is, to determine the type of plant.
    Incorrect identification of the type of plant could cause useful plants to be hoeed and count, measure and / or treat wild herbs.
    For the classification of plants, today we mainly use learning methods such as those described in the following documents "Plant classification System for crop / weed discrimination without segmentation Applications of Computer Vision (WACV)", 2014 IEEE Winter Conference 2014, 1142-1149; "Support Vector Machines for crop / weeds identification in maize fields Expert Systems with Applications", 2012, 39, 11149-11155 and "Evaluation of Features for Leaf Classification in Challenging Conditions Applications of Computer Vision (WACV)", 2015 IEEE Winter Conférence 2015 , 797-804 ”.
    To learn these classification units, a large volume of training and validation data is required, which must be provided with annotations. The annotations thus give the correct class of plants to be classified. For this reason, these annotations are called "basic truth or ground truth".
    All of the current procedures for obtaining truth from the field are very long and almost all are performed only manually. Despite the semi-automatic assistance by annotations such as for example the follow-up of annotations in several frames of the annotated video image, the semi-automatic segmentation of plants etc., the means to be implemented due to the quantity of annotations is very high because the means to be used for the set to be annotated are very high since the quantity of growth phases of useful plants and if possible of all possible weeds, must be preserved with the biological growth morphologies. Different parameters such as water, heat, soil, food drops, wind etc. involved in the growth and / or the condition of development of the plant.
    The means to be used for each image to be annotated increase more in crops than small intervals between plants, because of their overlap. Although the additional seizure of plants in the near infrared domain and by the application of the NDVI index (Differentiated Normalized Vegetation Index) makes it possible to segment the plant very effectively with respect to the background, it is complicated to arrive manually at dividing lines between plantations and weeds.
    Purpose of the invention
    Based on this state of the art, the present invention aims to develop a classification method for plants growing in a field as well as a classification unit applying the method, a control apparatus and a computer program for the implementation of the process.
    Presentation and advantages of the invention
    To this end, the subject of the invention is a method of classifying plants growing on a field consisting of:
    - enter the plants of a segment of the field using an optical and / or infrared input unit to obtain information on the respective plant image,
    - determine the analogy between the plants entered in the field segment, one in relation to the other and / or in relation to the plants of another field segment and / or of a reference element using the information of plant image, and
    - classify the plants seized according to the degree of analogy, the plant image information being assigned to at least one class.
    The invention also relates to a control device designed for:
    - determine the analogy between the plants entered in a segment of the field, between them and / or in relation to plants in another field segment and / or in relation to a reference element using the plant image information , the plants having been captured using an optical and / or infrared capture unit to obtain image information of the plant respectively, and
    - assign to the plants seized at least one class according to the degree of analogy with respect to the classification and the plant image information.
    The invention also relates to a computer program designed for:
    - determine the analogy between the plants entered in a field segment between them and / or in relation to the plants in another field segment and / or in relation to a reference element using the plant image information , the plants having been captured with an optical and / or infrared capture unit in order to obtain image information of the plant respectively, and
    - classify the plants seized according to a degree of analogy and assign the plant image information to at least one class.
    A field according to the invention is an agricultural surface, a plant cultivation surface or a plot or a segment of such a surface. The field is a cultivated field, a meadow or a pasture. The field can be an uncultivated field or a cultivated field or field. The field can also be covered or be part of a greenhouse. The plants to be classified are for example useful plants with fruits to be harvested, for example, agricultural products, fodder or plants used for their energy. The plants to be classified can also be decorative plants. Plants to be classified also include weeds.
    The field segment is an input segment or a segment of captured images or an image which can be captured by an optical input unit and / or by infrared. The field segment may include the entire captured image and / or a segment of the image captured by the optical and / or infrared capture unit. Thus, the field segment is an area in which we enter plants to classify. The field segment thus generally presents multiple plants.
    The input unit is an optical and / or multispectral and / or hyperspectral and / or infrared input unit. An optical capture unit is for example a camera or a 3D camera. The optical and / or infrared capture unit makes it possible to take images in the NIR domain and / or in the visible domain.
    The optical and / or infrared capture unit can be calibrated for example to calculate the height assigned from the captured images. The optical and / or infrared input unit can be installed on a mobile unit and this mobile unit is in particular an agricultural vehicle and / or an aerial vehicle and / or a trailer.
    Entering plants or a set of plants consists, for example, of determining the presence of plants in a field segment and / or of determining the shape, dimensions, species, number of leaves, leaf structure, the number of buds, the structure of the buds and other biological characteristics that distinguish plants from other objects or from the soil. Plants are entered in particular without classification of plants. By automatic segmentation processes such as, for example, processes using the NDVI index, it will be possible to segment the foreground relative to the ground and thus the plant, which makes it possible to grasp the plant. The NDVI index makes it possible to apply an efficient segmentation process but requires infrared color information. Other segmentation methods (for example methods based on RGB images) guarantee indices such as ExG (green in excess), ExG-ExR (green in excess red in excess). One can also envisage segmentations applying any other methods such as for example colors in different color spaces such as RGB, HSV, HSL, Lab, or relating to texture, or based on gradients.
    It is also conceivable that during the entry step, that is to say for example, when the mobile unit equipped with the optical and / or infrared entry unit passes, passing over the entry area, one or several plants are captured simultaneously in an image segment of the optical and / or infrared capture unit. We can also take several shots of the same plants from different perspectives during the passage to optimize the process. The unique image shooting technique also allows sequential shots of the same plant. This process reduces the influence of variable wind and light conditions, as well as image overlap and noise. Image information recorded by the optical and / or infrared capture unit under given lighting conditions includes, for example, RGB images and infrared images. The optical and / or infrared input unit is calibrated to allow for example to assign a height starting from the images. The images captured by the optical and / or infrared capture unit are recorded with each other, to allow a calculation for example in conjunction with the NDVI index (Differentiated Normalized Vegetation Index; This index is obtained from reflection values in the next infrared and in the visible red wavelengths of the spectrum). From the predefined NDVI index, we separate the biomass and the soil in the image, which results in one of the biomass masks (reduced image data with groupings of green components, which represent with great probability different plants).
    Plant image information is a representation parameter or a plant parameter or a plant characteristic which reproduces characteristics which can be entered optically or characteristics measurable with infrared radiation for the plant. Without departing from the scope of the invention, it is also possible in principle to use other wavelengths than the range of wavelengths of the visible domain and of the infrared (for example, multispectral or hyperspectral wavelengths ). Plant image information may also contain information obtained by processing the plant image captured by the optical camera and / or by an infrared sensor from the infrared capture unit. Plant image information can be learned by a control unit (for example by advanced learning). Plant image information can also be predefined. The plant image information is preferably plant characteristics and / or plant properties and / or plant states. Plant image information is preferably selected from the group comprising: shape information including 2D or 3D shape characteristics, gradients (for example, edges, corners, etc.), textures, colors (in different wavelength ranges and their derivatives), reflection coefficients (in different wavelength ranges and their derivatives), pixel statistics, intervals between plants, intervals between rows of plants, variation information, position information, development information, disease information. One can use or complete a position model, for example learned and / or parameterized with corresponding information.
    The determination unit is carried out to calculate the different entries in the field segment or partial surfaces of the image segment, the plant image information, that is to say the characteristics which describe the content of the plants. or partial surfaces. The exploitation can be done on the whole image or only on a frame, a sliding window (for example a frame with overlaps) and / or based on the object (for example from a segmentation of region attached).
    Plant image information is preferably saved when entered and / or classified in a memory unit.
    The analogy of the plants seized from the field segment between themselves and / or plants is determined with respect to those of another field segment and / or of a reference element by using information from images of plants for example. with a unit of determination. The determination unit can be part of a control unit and / or a control device. In the step of determining the analogy, the analogy coefficients of the plant image information are calculated; The analogy coefficients represent the degree of analogy between plants and / or with respect to a reference element. This plant image information can be compared according to known methods originating from unsupervised learning such as, for example, methods for detecting anomalies or grouping them together with the reference element. This allows us to see similar plants and those that are distinct. This also makes it possible to explicitly develop several classes or groups (for example with a proposal for a hierarchical or medium grouping k).
    The classification signifies an annotation, in particular an automatic annotation. Advantageously, in the classification step, plants whose plant image information have analogy coefficients within a predefined range and / or which can be predefined to a first class and / or plants are assigned the plant image information has analogy coefficients beyond the predefined range and / or which can be predefined, correspond to a second class.
    Thus, a class includes plants which have an identical or, at least, analogous property or correspond to an identical or at least analogous state such as for example nature, type, sort, unity of sort, age , size, state of health, level of growth, state of development. Thus, a class includes for example plants of a certain type of plant or a certain category of plants such as for example useful plants (sugar beet, corn, soybeans, wheat etc.) or weeds. The first class includes useful plants and the second class weeds. A class can also correspond to plants having other analogous characteristics such as size, state of development, state of health or other.
    The steps for entering plants, determining the analogy and classifying the plants seized can be carried out or controlled using a control unit. One can in particular carry out the steps of determining an analogy of classification of seized plants using an unattended learning process. The control unit can further be designed to filter specific features from the assigned or annotated plant image information and develop therefrom a model.
    In addition, provision may be made for treating plants classified according to the class. Treatment includes treatment with an agent such as, for example, a phytosanitary agent or mechanical treatment or destruction
    According to the invention, all the plants entered having a certain minimum degree of analogy or a determined minimum analogy are assigned at least plant image information or a plant characteristic, in a "flat-rate" manner of a class. common, for example the class "plant" without thereby carrying out the necessary identification of plants and annotation of plant image information with corresponding attributions to the different classes. As a variant or in addition, all the plants seized, remaining and which do not meet the analogy criterion, will also be assigned a "global" class, another class, for example, the class "foreign elements". Again, without doing the necessary identification and annotation of the plant image information.
    Thus, the method according to the invention allows without "prior knowledge" of plants by determining the analogy (by the method of unattended learning) to make a classification of plants in a field to recognize by one or more characteristics, analogous plants and / or plants which are distinguished from other plants (foreign elements) and classify them. This makes it possible to establish a classification without necessity of the fundamental data, that is to say, that all trained or learned classification units are totally avoided, which minimizes the means to be used for the classification of plants. This is used to collect by the method, in an advantageous manner, extremely large amounts of data including true information and true data and this with reduced means. They can be used to establish different classification units to characterize types of plants. Thanks to the large amount of basic data, it will be possible to separate the special classification units for any combination of useful plants distinct from weeds, types of plants of a certain growth size etc., by combining the information of plant images, classified or true data. There are considerable advantages in controlling weeds, in counting and measuring plants to characterize them by phenomenon, in selective harvesting as well as in the selective application of herbicide, fungicide, pesticide and insecticide.
    It is advantageous that according to another step of entering the plants of the other field segment be done using the optical and / or infrared input unit, before the step of entering the plants of the field segment for obtain plant image information respectively and this other field segment is a segment of the same field. Thanks to these means, other objects of comparison are obtained for the classification of new plants.
    In addition, it is advantageous to provide another step for classifying other plants using the plant image information assigned in particular by a trained classification unit. The trained or educated classification unit thus constitutes a long-term memory. The trained classification unit has been treated with true data and can distinguish the selected plant from another plant. The untrained classification unit can be considered as a short-term memory which contains only the history of a certain duration / quantity of image / plant and calculates from this information, the analogies in the case of a current picture. By the combination of the two units or processes, the classification unit trained during operation can be trained a posteriori and thus adapted in the long term according to requirements. Consequently, the robustness is improved, for example for variable conditions (new seed, other lighting conditions, other entry conditions, wind, time of day, humidity, etc.).
    The solution according to the invention also develops a classification unit for carrying out the steps according to a variant of the process presented in appropriate facilities. This variation in the embodiment of the invention in the form of a unit or a device makes it possible to quickly and effectively solve the problem of the invention.
    The classification unit comprises for this purpose at least one calculation unit for processing the signals or the data, at least one memory unit for recording the signals or the data, at least one interface with a sensor or an actuator for recording the sensor signals supplied by the sensor or for transmitting data or command signals to the actuator and / or at least one communication interface for receiving or transmitting data integrated in a communication protocol. The computing unit is for example a signal processor, a microcontroller or the like and the memory unit is a flash memory, an EPROM memory or a magnetic memory. The communication interface makes it possible to record and supply data by a wireless link and / or a cable link and the communication interface makes it possible to receive or send data by a line, this data being received by for example by electrical or optical means from a corresponding data transmission line or be supplied by a corresponding data transmission line.
    A classification unit according to the invention is an electrical device which processes the signals of sensors and supplies therefrom control signals and / or data signals. The device can include an interface in the form of a wired embodiment and / or a program. In the case of a wired embodiment, that is to say in the form of a circuit, the interfaces are for example a part of an ASIC system which contains the most diverse operations of the device. It is also possible that the interface has its own integrated circuit or at least that the interface is made in the form of discrete components. In the case of implementation in the form of a program, the interfaces are program modules which are, for example, housed in the microcontroller next to other program modules. It is also advantageous to carry out the invention in the form of a computer program product or more simply of a computer program with a program code recorded on a medium readable by a machine or a memory medium such as a semiconductor memory, a hard disk or an optical memory for the execution, the application and the control of the process steps, according to any of the embodiments described above, in particular when the product program or more simply the program is executed by a computer or a device. The method according to the invention can for example be implemented in the form of a program or a circuit or in a mixed form between program and circuit by a control device.
    drawings
    This will be described below, in more detail using an example of a method and a classification unit, shown in the accompanying drawings in which:
    - Figure 1 is the diagram of a classification unit according to the invention,
    - Figure 2 shows a simplified flowchart of a method of an exemplary embodiment of the classification method according to the invention.
    Description of embodiments
    FIG. 1 shows a classification unit according to the invention, generally given the reference 10.
    The classification unit 10 is installed on a carriage 12. The classification unit 10 is an untrained, that is to say, untrained classification unit 10. The classification unit 10 has an input unit 14 and a control unit 16.
    The input unit 14 is produced in the form of an optical input unit 14 or an optical camera 14. The optical camera 14 includes a filter unit 18 for extracting a color component such as for example the green component and / or the red component of an image of a field segment 20 captured by the optical camera 14 and corresponding to a field 22 for capturing plants 24 to be classified.
    The control unit 16 further comprises a calculation unit 26 and a memory unit 28. The control unit 16 filters the information of the plant image of the plant 24 captured by the camera 14 and saves it. in the memory unit 28. The control unit 16 is further adapted to determine or calculate the analogy or a coefficient of analogy of the plants seized 24 in the field segment 20 between them using the image information of plants. The control unit 16 is further adapted to classify the captured plants 24 according to the degree of analogy and to assign the plant image information to at least one class.
    The carriage 12 also includes a device 30 for controlling weeds. This weed control device 30 comprises a driven classification unit 32 and a spraying unit 34. The weed control device 30 cooperates with the control unit 16.
    Thus, the classification unit 32, trained, which informed with true data can distinguish a selected plant from another plant, is permanently informed with the classified plants 24, that is to say with information of assigned plant image; it is thus adapted in the long term according to variable conditions / data. Using the trained classification unit 32 (post-trained), it will be possible to better classify other plants 24 and treat or control them correspondingly using the spraying unit 34.
    FIG. 2 shows a simplified flowchart of an exemplary embodiment of the example of method 100 of classification of plants 24 in a field 22. The method 100 comprises a step 104 of entering plants 24 of the segment 20 of field 22 to 1 using an optical or infrared input unit 14 to obtain plant image information 24 respectively. The method 100 further comprises a step 106 of determining the analogy between the plants captured 24 of the field segment 20, relative to each other and / or to plants of another field segment and / or a reference element using the plant image information. The method 100 also includes a step 108 of classification of the plants 24 seized according to a degree of analogy; the plant image information is thus assigned to at least one class.
    Optionally, the method 100 can also include a step 102 for entering plants 24 'from another field segment 20' using the optical and / or infrared input unit 14 before step 104 for capturing the plants 24 of the field segment 20 to have respectively plant image information 24 ', this other segment 20' being a segment of the field 22.
    Optionally, the method 100 also includes a step
    110 of classification of other plants 24 using the information 5 of plant images, allocated in particular using the classification unit 30, informed.
    权利要求:
    Claims (13)
    [1" id="c-fr-0001]
    1) Method (100) for classifying plants (24) growing on a field (22), method comprising the following steps consisting in:
    - enter (104) the plants (24) of a segment (20) of the field (22) using an optical and / or infrared capture unit (14) to obtain respective image information of the plant (24),
    - determining (106) the analogy between the plants entered (24) in the field segment (20), with respect to each other and / or with respect to the plants (24 ') of another field segment (20 ') and / or a reference item using the plant image information, and
    - classifying (108) the plants seized (24) according to the degree of analogy, the plant image information being assigned to at least one class.
    [2" id="c-fr-0002]
    2 °) Method (100) according to claim 1, characterized in that in the step of determining the analogy (106), we calculate the coefficient of analogy of the plant image information (24, 24 ') ,
    - the analogy coefficients representing the degree of analogy between the plants (24, 24 ’) and / or the reference element.
    [3" id="c-fr-0003]
    3 °) Method (100) according to claim 1 or 2, characterized in that
    - in the classification step (108), a first class is assigned to the plants (24) whose plant image information have analogy coefficients within a predefined range and / or likely to be,
    - And / or we assign to a second class, plants (24) whose plant image information have analog coefficients outside the predefined and / or predetermined range.
    [4" id="c-fr-0004]
    4 °) Method (100) according to one of the preceding claims, characterized in that the steps (106, 108) of determination and classification are carried out using an unattended learning process.
    [5" id="c-fr-0005]
    5 °) Method (100) according to one of the preceding claims, characterized in that the plant image information (24, 24 ') is selected from the group comprising: image information in particular 2D shape characteristics or 3D, color values, reflection coefficients, pixel statistics, plant spreads, plant treatment residues, position information, variation information, development information, disease information.
    [6" id="c-fr-0006]
    6 °) Method (100) according to one of the preceding claims, characterized by the following step consisting in:
    - enter (102) the plants (24 ') of the other field segment (20') using the optical and / or infrared input unit (14) before the input step (104) plants (24) of the field segment (20) to obtain respectively information of images of the plant (24 '),
    - this other field segment (20 ’) being a field segment (22).
    [7" id="c-fr-0007]
    7 °) Method (100) according to one of the preceding claims, characterized by the other step consisting in:
    - classify (110) the other plant (24) using the allocated plant image information in particular using the trained classification unit (32).
    [8" id="c-fr-0008]
    8 °) Method (100) according to one of the preceding claims, characterized in that in the entry step (104), the plants (24) are entered using a color value in particular the component of green color and / or the red component and / or the infrared component.
    [9" id="c-fr-0009]
    9 °) Method (100) according to one of the preceding claims, characterized in that the step of entering (104) plants (24) is done using a mobile unit (12) equipped with the optical and / or infrared input unit (14) which is in the form of a vehicle and / or an airplane.
    [10" id="c-fr-0010]
    10 °) Classification unit (10) designed to apply the method (100) according to one of the preceding claims with the units (14, 16) consisting of:
    - enter (104) the plants (24) of a segment (20) of the field (22) using an optical and / or infrared capture unit (14) to obtain respective image information of the plant (24),
    - determining (106) the analogy between the plants entered (24) in the field segment (20), with respect to each other and / or with respect to the plants (24 ') of another field segment (20 ') and / or a reference item using the plant image information, and
    - classifying (108) the plants seized (24) according to the degree of analogy, the plant image information being assigned to at least one class.
    [11" id="c-fr-0011]
    11 °) Control device designed for:
    - determine the analogy between the plants seized (24) of a segment (20) of the field (22) between them and / or with respect to the plants (24 ') of another field segment (20') and / or relative to a reference element using the plant image information (24),
    - the plants (24) having been captured using an optical and / or infrared capture unit (14) to obtain image information of the plant (24) respectively, and
    - assign to the plants seized (24) at least one class according to the degree of analogy with respect to the classification and the information on plant images.
    [12" id="c-fr-0012]
    12 °) Computer program designed for:
    - determine the analogy between the plants seized (24) of a segment (20) of the field (22) between them and / or with respect to the plants (24d of another segment of field (20d and / or with respect to an element of re-
    5 reference using plant image information (24),
    - the plants (24) having been captured with an optical and / or infrared capture unit (14) to obtain plant image information (24) respectively, and
    - classify the plants seized (24) according to a degree of analogy and
    10 assign the plant image information to at least one class.
    [13" id="c-fr-0013]
    13 °) Memory medium readable by a machine containing the computer program according to claim 12.
    类似技术:
    公开号 | 公开日 | 专利标题
    Hamuda et al.2016|A survey of image processing techniques for plant extraction and segmentation in the field
    Annamalai et al.2004|Color vision system for estimating citrus yield in real-time
    Pan et al.2008|Recognition of plants by leaves digital image and neural network
    Esau et al.2018|Machine vision smart sprayer for spot-application of agrochemical in wild blueberry fields
    FR3071644A1|2019-03-29|METHOD AND DEVICE FOR CLASSIFYING PLANTS
    Diago et al.2016|Assessment of vineyard canopy porosity using machine vision
    Williams et al.2017|A method for automatic segmentation and splitting of hyperspectral images of raspberry plants collected in field conditions
    Yano et al.2016|Identification of weeds in sugarcane fields through images taken by UAV and random forest classifier
    Diago et al.2019|On‐the‐go assessment of vineyard canopy porosity, bunch and leaf exposure by image analysis
    Calou et al.2020|The use of UAVs in monitoring yellow sigatoka in banana
    Parra et al.2020|Edge detection for weed recognition in lawns
    De Smedt et al.2011|Neural networks and low-cost optical filters for plant segmentation
    Pforte et al.2012|Comparison of two different measurement techniques for automated determination of plum tree canopy cover
    Sood et al.2021|Computer Vision and Machine Learning based approaches for Food Security: A Review
    Muppala et al.2020|Machine vision detection of pests, diseases and weeds: A review
    Khandelwal et al.2019|Artificial Intelligence in Agriculture: An Emerging Era of Research
    Okamoto et al.2008|Unified hyperspectral imaging methodology for agricultural sensing using software framework
    Rasti et al.2021|A survey of high resolution image processing techniques for cereal crop growth monitoring
    EP3579186A1|2019-12-11|Method and system for managing an agricultural plot
    Tripathy et al.2020|Image processing techniques aiding smart agriculture
    Su et al.2022|Spectral analysis and mapping of blackgrass weed by leveraging machine learning and UAV multispectral imagery
    Amaral et al.2021|UAV applications in Agriculture 4.0
    Negrete2018|Artificial Vision in Mexican Agriculture for Identification of diseases, pests and invasive plants
    Zheng et al.2021|Phenoliner2. 0: RGB and near-infrared | image acquisition for an efficient phenotyping in grapevine research
    Ruett et al.2022|Hyperspectral imaging for high-throughput vitality monitoring in ornamental plant production
    同族专利:
    公开号 | 公开日
    FR3071644B1|2021-12-03|
    DE102017217258A1|2019-03-28|
    引用文献:
    公开号 | 申请日 | 公开日 | 申请人 | 专利标题
    JPH0158548B2|1981-04-27|1989-12-12|Tokyo Shibaura Electric Co|
    DE102009023896B4|2009-06-04|2015-06-18|Fraunhofer-Gesellschaft zur Förderung der angewandten Forschung e.V.|Apparatus and method for detecting a plant|
    DE102015221085A1|2015-10-28|2017-05-04|Robert Bosch Gmbh|Method and information system for recognizing at least one plant planted in a field|DE102019211642A1|2019-08-02|2021-02-04|Robert Bosch Gmbh|Method for identifying weeds within a defined row of plants in an agricultural area|
    DE102019218186A1|2019-11-25|2021-05-27|Robert Bosch Gmbh|Method of working crops in a field|
    法律状态:
    2019-09-23| PLFP| Fee payment|Year of fee payment: 2 |
    2020-09-21| PLFP| Fee payment|Year of fee payment: 3 |
    2020-12-11| PLSC| Publication of the preliminary search report|Effective date: 20201211 |
    2021-09-27| PLFP| Fee payment|Year of fee payment: 4 |
    优先权:
    申请号 | 申请日 | 专利标题
    DE102017217258.4A|DE102017217258A1|2017-09-28|2017-09-28|Method for classifying plants|
    DE102017217258.4|2017-09-28|
    [返回顶部]