• 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:
    The invention relates to a method of recognizing humans by analyzing a detected object (P) in a monitored area and whether or not the detected object is a human being with a laser scanner (12), comprising : the laser scanner (12) generates at least one laser curtain (22, 32, 34), each laser curtain (22, 32, 34) being generated by multiple pulses evaluated by pulse time (TOF) measurement of pulses individual to generate the distance of the reflection points from the position of the laser scanner; a combination of distances of the reflection points with the direction of the pulse to recover a position in a predefined detection zone within a monitored region; project the reflection points belonging to a detected object into an evaluation plan (EP) as evaluation objects (O1, O2), the evaluation plan (EP) having a Z axis which is associated with the height and an axis perpendicular to the Z axis which is associated with the width in the direction of the lateral extension of the laser curtain (22, 32, 34). The invention is characterized in that the evaluation plan (EP) is evaluated on the basis of the density distribution of the points of reflection along the Z axis and the evaluation result is compared with anthropometric parameters. .
    公开号:BE1025329A9
    申请号:E20185228
    申请日:2018-04-05
    公开日:2019-02-05
    发明作者:Gautier Radermecker
    申请人:Bea Sa；
    IPC主号:
    专利说明:

    Human body recognition method and human body recognition sensor
    A method for recognizing a human body and a human recognition sensor for detecting an object in a monitored region and deciding whether or not the detected object is a human body are provided.
    EP 2,741,263 B1 discloses a human recognition sensor for intrusion detection, which comprises a distance acquisition unit for defining reflection points of a detected object and a human recognition unit for defining if a detected object is a human body, based on an estimated height and width of an object.
    The objective of the invention is to improve the accuracy of the human detection sensor for control purposes.
    A method of analyzing a detected object within a monitored region and deciding whether or not the detected object is a human body with a laser scanner, comprising the steps of: the scanner The laser generates at least one laser curtain, each laser curtain being generated by multiple pulses evaluated by measuring the flight time of individual pulses to generate the distance of the reflection points from the position of the laser scanner. In addition, a combination of distances from the reflection points with the direction of the pulse is achieved to recover a position in a predefined detection area within a monitored region. The position recovered from the reflection points belonging to a detected object is projected into an evaluation plane having a Z axis which is associated with the height and an axis perpendicular to the Z axis which is associated with the width in the direction of l lateral extension of the laser curtain.
    According to the invention, the evaluation plan is evaluated on the basis of the density distribution of the reflection points on the Z axis and the evaluation result is compared with anthropometric parameters.
    2018/5228
    BE2018 / 5228 2
    The monitored region is defined by the laser curtain and has a vertical height direction and two lateral directions, a depth and a width, which are all perpendicular to each other. In the case of a single vertical laser curtain, the depth of the monitored region is equal to the depth of the laser curtain.
    The evaluation plane can have a Z axis which corresponds to the vertical axis of the vertical plane and / or an evaluation width extension which corresponds to the width of the region monitored. However, the Z axis, for example, can be defined along a laser curtain inclined to the vertical direction, but the width can still correspond to the width of the laser curtain.
    Anthropometric parameters according to the invention are measurements of the human body and / or proportions of the human body.
    In particular, anthropometric parameters are parameters which relate in particular to the height, the width, the shoulder width, the shoulder height, the head width, the total height of a human body.
    On the basis of the density distribution in the evaluation plan, the evaluation unit decides whether or not the density distribution corresponds to that of a human body.
    To determine if a detected object is a human body, the density distribution along the Z axis is evaluated, the Z axis representing the height of a detected object. The density distribution corresponding to a human body includes two peaks, one peak being approximately at the top of the head and the second peak being approximately at the top of the shoulder.
    The determination is preferably made to determine the ratio of the height of the head to the height of the shoulder. Since the ratio head height to shoulder height is an anthropometric parameter which is essentially equal for all human beings and above all it is not dependent on absolute height, a reliable distinction of human beings is possible according to the density distribution assessment.
    In addition to the density distribution, the evaluation unit can evaluate the width of an object in an additional step. Therefore, it analyzes the points of reflection in the plan
    2018/5228
    BE2018 / 5228 3 evaluation belonging to an object at the position of the density distribution peaks and determines the effective head and shoulder width of the human body.
    Due to the integration of this information, the assessment can be obtained more precisely. A valid head width to shoulder width ratio can be predefined to check if it corresponds to the result deduced from the evaluation density distribution evaluation. The result can be compared to the result of the density evaluation. If both assessments are positive, it is most likely that the object detected is a human body.
    In addition, the evaluation unit can count the number of reflection points within the peak areas of the density distribution evaluation. If the number is below a predefined number, the measurement will be ignored.
    The movement of the human body takes place in a direction of movement, the direction of movement being essentially a vector of width and depth. In particular, in door applications, the direction of movement is perpendicular to the direction of width and, therefore, the orientation of the shoulders of a human body is usually aligned with the direction of width.
    According to the invention, individual evaluation objects can be identified among all the reflection points of the evaluation plan and a subset of reflection points is created for each evaluation object, which is then subjected to an analysis. density distribution.
    Depending on this, there may be a decision on each evaluation object present as to whether or not it corresponds to a human body. As a result, detection sensors can base their decision to control doors or lights on whether a detected object is a human body or not.
    The determination of individual evaluation objects is carried out by the evaluation unit, the evaluation plan, containing all the reflection points, being analyzed by a neighboring zone, from the top to the bottom of the plan. Once a reflection point or reflection points are newly present in the neighboring area, all reflection points within the neighboring area are taken into account and the newly present reflection point is assigned to a object
    2018/5228
    BE2018 / 5228 4 evaluation. It is assigned to a new evaluation object if there is not another point on the newly present point inside the neighboring area, or to an existing evaluation object if the reflection point has the smallest distance to the mathematical center of gravity of an existing evaluation object.
    According to this procedure, all the reflection points are grouped in a subset of reflection points belonging to an evaluation object.
    According to this assessment, even two or more people walking parallel through the laser curtain can be distinguished.
    According to another improvement of the invention, the reflection points can be integrated in time on the evaluation plan. This leads to a higher density of reflection points and, therefore, evaluation objects can be better distinguished and detected objects can be classified more reliably.
    The time integration can be performed on the basis of a fixed time interval after a first detection of a detected object has taken place.
    According to another improvement of the invention, integration over time is carried out in such a way that the subset of reflection points is assigned to a temporal object by projection of the reflection points in a width-time plane, the height of the reflection point being ignored. The width axis extends as a function of a predefined accumulation / integration time.
    The reflection points projected in the time-width plane are grouped into subsets assigned to temporal objects. Each time object is the main set of reflection points to generate the evaluation plan, the time component of the reflection point being neglected but the height being taken into account.
    According to this procedure, a more precise decision on the delimitation of temporal objects is possible. Consequently, the information acquired is more precise with regard to the number of humans passing by.
    2018/5228
    BE2018 / 5228
    The grouping of time objects is preferably carried out using a spatial grouping algorithm by density of applications in the presence of noise (DBSCAN).
    Preferably, the scanner generates multiple laser curtains which are inclined with respect to each other. Due to the different laser curtains, a more precise image can be taken and the direction of movement of the object can be taken into account.
    The scanner, preferably, evaluates and / or generates multiple laser curtains in succession.
    By taking into account at least two curtains, which are inclined with respect to each other, two depth positions perpendicular to the width of the scanning plane can be evaluated. As the two planes are scanned in succession, the direction of movement of a human being can be detected since the center of gravity during a scanning time changes in the time-width diagram in the direction of movement of the detected object.
    When using multiple laser curtains, a predefined accumulation time for time integration is longer or equal to the time it takes to scan the laser curtains present from the sensor.
    The evaluation unit may not accept reflection points which clearly refer to background effects. Therefore, background noise can be reduced at this point.
    The invention further relates to a human recognition sensor for analyzing an object in a monitored region and deciding whether or not the object is a human body, comprising a laser scanner and an evaluation unit which is capable of performing a process as described above.
    Another aspect relates to a sensor which generates at least one laser curtain which is inclined by less than 45 ° relative to the vertical axis. This allows aerial scanning so that human bodies can be recognized as they pass under the sensor.
    The human recognition sensor may comprise a computer unit, preferably a microprocessor, a microcontroller or a programmable pre-broadcast network (FPGA), on which
    2018/5228
    BE2018 / 5228 6 the evaluation unit is implemented in the form of a software program, executing the method described above.
    Other characteristics, advantages and potential applications of the present invention can be gathered from the description which follows, together with the embodiments illustrated in the drawings.
    Throughout the description, the claims and the drawings, the associated terms and reference numbers will be used as noted in the attached list of reference numbers. In the drawings are represented
    Fig. 1 a schematic view of a laser scanner according to the invention; Fig. 2 a first embodiment of a human recognition sensor having a scanning curtain; Fig. 3 a method of human recognition by a sensor of FIG. 1; Fig. 4 a second embodiment of a human recognition sensor having two scanning curtains; Fig. 5a an operating principle of the evaluation unit describing a first step by generation of time objects; Fig. 5b an enlarged view of a created temporal object; Fig. 6a a view of the temporal object of FIG. 4b in the evaluation plan; Fig. 6b a view of the temporal object after separation of human objects; Fig. 7a a human object separated from FIG. 5b; Fig. 7b a density distribution of the human object of FIG. 6a;
    2018/5228
    BE2018 / 5228 7
    Fig. 8 has a time-width view on the time object of FIG. 4b for the first sweep curtain, and
    Fig. 8b a time-width view for the time object of FIG. 4b for the second curtain.
    Fig. 1 shows a first embodiment of the human recognition sensor 10 according to the invention. The human recognition sensor 10 comprises a laser scanner 12, a processing unit 14, the processing unit 14 comprising an evaluation unit 16. The processing unit 14 is connected to the laser scanner 12 as well as to an output port 18 into which information can be injected which contains information relating to human recognition results.
    The laser scanner of the embodiment according to FIG. 1 uses at least one laser curtain which is evaluated by taking into account the point of the reflections which are deduced by light pulses (where the time of flight (TOF) is determined. According to this determination of time of flight and the direction of the pulse, a position of the reflection point with respect to the laser scanner can be deduced This evaluation can be carried out by the processing unit 14, relevant reflection points being determined and their position being injected into the unit d evaluation 16 which performs the method according to the invention as described in more detail in conjunction with the figures which follow.
    According to this configuration, the evaluation unit 16 receives the data from the reflection point with regard to the laser scanner 12.
    The evaluation unit 16 then analyzes the point of the reflections according to the invention as will be described in more detail in the following figures. And, accordingly, it will output a signal containing information as to whether or not a detected object is a human body.
    Fig. 2 shows an example of application where the human recognition sensor 20 is mounted in a high position; there are objects passing under it. The human recognition sensor 20 projects a laser curtain which extends in a vertical direction
    2018/5228
    BE2018 / 5228 8 in a direction of width W. It is represented while a person P moves through the laser curtain 22 in a direction of movement M. The person P passing by reflects pulses of light, the laser scanner of the sensor of human recognition 20 evaluating the point of reflection inside the laser curtain.
    The evaluation unit of the sensor 20 is adjusted in such a way that it evaluates an evaluation plane EP which corresponds to the laser curtain 22. Consequently, the evaluation plane EP has a Z axis in a vertical direction and the same axis of width W as the laser curtain 22.
    Fig. 3 shows the process for recognizing the human body by evaluating the EP evaluation plan, in this case the points of reflection not having to be projected into the EP evaluation plan since the EP evaluation plan corresponds to the laser curtain 22. The reflection points are applied to the EP evaluation plan according to their position. The evaluation plan EP has an axis Z and an axis of width W.
    According to the invention, the evaluation unit 16 now calculates a density distribution along the axis Z of the evaluation plane EP, in this density distribution two peaks being assumed to be deducible.
    If there is, for example, only one peak, the measurement is discarded and the evaluation object is not identified as a human body.
    If there are two peaks 24, 26, as would be the case when detecting a human body, the position H1, H2 of the position of the peaks on the Z axis is taken. The first peak 24 is assumed to provide the total height H1 of the object, which is the head when looking at the human body, and the second peak 26 is assumed to be the shoulder height H2 of a person. The ratio of total height H1 to shoulder height H2 is compared to a range of predefined human body proportions. In addition, head height (the distance between shoulder height and total height; H1-H2) can also be taken into account, since the proportions of the human body change with the age of human beings.
    2018/5228
    BE2018 / 5228 9
    According to this, it is not necessary to limit the measurement to a minimum height which could possibly exclude children from detection, since they can be defined according to the evaluation described above.
    Within the evaluation plan EP, the width W2 of the shoulders and the position H2 of the second density peak 26 can be determined. In the region of the first peak 24, the width of the head W1 can be determined. Due to these additional parameters, a more precise evaluation of the object with regard to recognition of the human body can be obtained.
    Fig. 4 shows a configuration with a human recognition sensor 30 which generates multiple laser curtains 32, 34. The human recognition sensor 30 in this case is mounted above the door frame and monitors the region in front of the door . The laser curtains 32, 34 are inclined with respect to the vertical axis and one with respect to the other and extend parallel to the door in a direction of width W. The evaluation plane EP is defined parallel to the door plan.
    The laser scanner of the human recognition sensor 30 deduces the position of the reflection points of the detected object relative to the laser scanner, the evaluation unit projecting them into the evaluation plan EP as objects of 'Evaluation.
    People P, as they move through the laser curtains 32, 34, produce reflection points during an acquisition period.
    As depicted in FIG. 5a, the acquisition period is approximately 15 seconds. In the case described, four detected objects pass in succession through the laser curtains, two detected objects passing through the laser curtains at the same time. The evaluation unit is implemented to project the points of reflection acquired in a time-width plane.
    In this time-width plane, the reflection points present are grouped into time objects TO_1, TO_2, TO_3. This is done using the spatial grouping algorithm by density of applications in the presence of noise (DBSCAN).
    2018/5228
    BE2018 / 5228 10
    The four objects detected passing the laser curtain during the acquisition period in this case lead to the definition of three time objects TO_1, TO_2, TO_3.
    As shown in an enlarged view of the time object TO_2, there may be more objects detected in the time object TO_2.
    The evaluation unit is furthermore arranged to take the reflection points of each time object and project them into the evaluation plan EP, as shown in FIG. 6a. The evaluation plan has a vertical axis Z and an axis of width W.
    In a following separation step, the evaluation unit assigns the reflection points of each time object TO_1, TO_2, TO_3 to objects.
    This is done by analyzing the EP evaluation plan from top to bottom and by assigning each point to an evaluation object.
    The individual evaluation objects O1 are determined by the evaluation unit, the evaluation plan EP containing all the reflection points of the time object TO_2. The EP evaluation plan is analyzed by a neighboring zone 40 from the top to the bottom of the EP evaluation plan. Once a reflection point or reflection points are newly present in the neighboring area 40, all the reflection points inside the neighboring area 40 are taken into account and the newly present reflection point is assigned to an evaluation object; for example see in Fig. 6b an O2 object (cross) and an O1 object (circles). It is assigned to a new evaluation object, if there is not another point on the newly present point inside the neighboring area, or to an existing evaluation object when the reflection point has the smallest distance to the mathematical center of gravity of an existing object, O1 or O2. According to this procedure, all the reflection points are grouped in a subset of reflection points belonging to an evaluation object O1, O2.
    Consequently, FIG. 6b shows that the time object TO_2 of FIG. 5b has been separated into two evaluation objects O1, O2.
    2018/5228
    BE2018 / 5228 11
    Each object in this evaluation plan, as shown in Fig. 7a, is then subjected to the density distribution analysis along the Z axis, as shown in FIG. 7b. In Figs. 7a, 7b, the object O1 is analyzed. The additional assessment to determine whether or not an object is a human body is performed as described in Fig. 3, by comparison of the measurements applied to anthropometric data.
    According to another improvement of the invention, the evaluation unit may be able to analyze the direction of movement of objects. This allows the human recognition sensor to provide direction information with the object information. For example, this allows a count of how many people enter or leave a building or totals itself and simply provides the net count at the port of exit.
    The direction of movement is analyzed by comparison of the reflection points accumulated from the two curtains 32, 34 over a short period of time, for example 500 ms. The reflection points are projected into a time-width plane, in which the mathematical center of gravity of the reflection points present is determined for each curtain.
    According to the displacement of the center of gravity, indicated by the cross in Fig. 8a and in FIG. 8b, the center of gravity passes first through the first curtain 32, then through the second curtain 34, which is then the direction of movement of the object.
    2018/5228
    BE2018 / 5228
    List of reference figures
    10 human recognition sensor 12 laser scanner 14 computer unit 16 evaluation unit 18 output port 20 human recognition sensor 22 laser curtain 24 peak 26 peak 30 human recognition sensor 32 first laser curtain 34 second laser curtain 44 center of gravity 46 center of gravity TO_1 time object TO_2 time object TO_3 time object O1 object of evaluation O2 object of evaluation EP evaluation plan P person M direction of travel Z Z axis W width axis
    2018/5228
    BE2018 / 5228
    Figure 3
    Z axis
    Width (m)
    Z axis
    Density (% of number of points)
    Figure 5a
    Time (s)
    Width (m)
    Figures 6a, 6b, 7a
    Height (m)
    Width (m)
    Figure 7b
    Height (m)
    Density (% of number of points)
    Figure 8a
    Curtain 1
    Time (s)
    Width (m)
    Figure 8b
    Curtain 2
    Time (s)
    Width (m)
    权利要求:
    Claims (15)
    [1]
    1. A method of recognizing a human body by analyzing a detected object (P) in a monitored region and by decision as to whether or not the detected object is a human being with a laser curtain (12), comprising:
    the laser scanner (12) generates at least one laser curtain (22, 32, 34), each laser curtain (22, 32,
    34) being generated by multiple pulses evaluated by time of flight measurement (TOF) of individual pulses to generate the distance of the reflection points from the position of the laser scanner;
    a combination of distances from the reflection points with the direction of the pulse to recover a position in a predefined detection area within a monitored region; project the reflection points belonging to an object detected in an evaluation plan (EP) as evaluation objects (O1, O2), the evaluation plan (EP) having a Z axis which is associated with the height and an axis perpendicular to the Z axis which is associated with the width in the direction of the lateral extension of the laser curtain (22, 32, 34), characterized in that the evaluation plane (EP) is evaluated based on the density distribution of the reflection points along the Z axis and the evaluation result is compared to anthropometric parameters.
    [2]
    2. Method according to claim 1, characterized in that the anthropometric parameters are measurements of the human body and / or proportions of the human body.
    [3]
    3. Method according to claim 1 or 2, characterized in that the reflection points belonging to an evaluation object (O1, O2) are evaluated on the basis of a density distribution over the height, from which are deducted therefore a head height (H1) and a shoulder height (H2), and the anthropometric parameter is the ratio head height (H1) to shoulder height (H2), which is compared to a predefined range for a human body.
    [4]
    4. Method according to any one of the preceding claims, characterized in that the head height (H1) and the shoulder height (H2) are deduced by evaluation of the peaks (24, 26) of the density distribution.
    [5]
    5. Method according to any one of the preceding claims, characterized in that the evaluation plane (EP) is evaluated due to the density distribution over the height, a
    2018/5228
    BE2018 / 5228 15 head width (W1) and a shoulder width (W2) being deduced by taking the width (W1, W2) at the peaks of the corresponding density distribution.
    [6]
    6. Method according to claim 5, characterized in that the parameter
    5 Anthropometric is the ratio head width (W1) to shoulder width (W2), which is compared to a predefined range for a proportion of human body.
    [7]
    7. Method according to any one of the preceding claims, characterized in that the sensor generates and evaluates a plurality of laser curtains (22, 32, 34) which are inclined the
    10 relative to each other.
    [8]
    8. Method according to claim 7, characterized in that the various laser curtains (32, 34) are swept in succession and measure the direction of movement (M) of a human body and / or of an object analyzed.
    [9]
    9. Method according to any one of the preceding claims, characterized in that the reflection points are integrated in time over an acquisition period.
    [10]
    10. Method according to any one of the preceding claims 9, characterized by the
    20 causes a subset of reflection points to be assigned to temporal objects (TO_1, TO_2, TO_3) which are defined by grouping reflection points in a width-time plane over the acquisition period.
    [11]
    11. Method according to claim 10, characterized in that the subset of
    25 points of reflection assigned to the temporal object (TO_1, TO_2, TO_3) is the main set for an evaluation on the evaluation plan (EP).
    [12]
    12. Method according to any one of the preceding claims, characterized in that a subset of reflection points in the evaluation plan (EP) is assigned to an object
    30 evaluation (O1, O2) by grouping reflection points.
    [13]
    13. The method of claim 12, characterized in that the reflection points are analyzed with respect to the criterion of whether there is another point above the point
    2018/5228
    BE2018 / 5228 16 currently analyzed in the Z direction within a certain range, if there is none, the point is assigned to a new object, if there are points, the point is assigned to an existing object which is the one having the center of gravity.
    5
    [14]
    14. Human recognition sensor for detecting and analyzing a detected object (P) in a monitored region and deciding whether or not the detected object is a human being, comprising a laser scanner (12) and an evaluation unit (16) which is capable of carrying out a method according to any one of the preceding claims.
    15. Human recognition sensor according to claim 14, characterized in that the evaluation unit (16) is a computer unit, for example a microprocessor, and the method is a method implemented by computer.
    16. Human recognition sensor according to any one of the claims
    [15]
    15 above 14 to 15, characterized in that the at least one laser curtain (22, 32, 34) is inclined by less than 45 ° relative to the vertical axis.
    类似技术:
    公开号 | 公开日 | 专利标题
    EP3087341B1|2019-08-28|Method for reconstructing a surface using spatially structured light and a dynamic vision sensor
    Parrish et al.2011|Empirical comparison of full-waveform lidar algorithms
    BE1025329A9|2019-02-05|Human body recognition method and human body recognition sensor
    US11243294B2|2022-02-08|Waveform reconstruction in a time-of-flight sensor
    JP6309099B2|2018-04-11|Monitoring device
    CN106037651A|2016-10-26|Heart rate detection method and system
    EP0237554B1|1992-03-25|Process and device for measuring the stare direction
    CN102538695B|2014-03-12|Security detection method and related device thereof
    BE1025335B1|2019-01-29|Sensor to control an automatic door
    Elseberg et al.2011|Full wave analysis in 3D laser scans for vegetation detection in urban environments
    FR3030056B1|2019-10-11|METHOD FOR DETECTING MOBILE ELEMENTS IN A BUILDING AND INSTALLATION FOR IMPLEMENTING SAID METHOD
    US8805075B2|2014-08-12|Method and apparatus for identifying a vibrometry spectrum in imaging applications
    US10999524B1|2021-05-04|Temporal high dynamic range imaging using time-of-flight cameras
    Kulikov et al.2012|Method of abandoned object detection based on a sequence of depth images
    JP2019128221A|2019-08-01|Time measurement device, distance measurement device, moving body device, time measurement method, and distance measurement method
    JP5743635B2|2015-07-01|Foreign object detection device
    同族专利:
    公开号 | 公开日
    BR112019021101A2|2020-05-12|
    SG11201909259PA|2019-11-28|
    EP3610291A1|2020-02-19|
    KR20190132508A|2019-11-27|
    BE1025329A1|2019-01-23|
    US20210011126A1|2021-01-14|
    EP3388864A1|2018-10-17|
    AU2018251941A1|2019-11-21|
    CA3059267A1|2018-10-18|
    BE1025329B1|2019-01-29|
    WO2018189191A1|2018-10-18|
    CN111033314A|2020-04-17|
    JP2020516910A|2020-06-11|
    引用文献:
    公开号 | 申请日 | 公开日 | 申请人 | 专利标题
    JP2006236184A|2005-02-28|2006-09-07|Nec Engineering Ltd|Human body detection method by image processing|
    US8611604B2|2009-06-03|2013-12-17|Chubu University Educational Foundation|Object detection device|
    JP5369036B2|2010-03-26|2013-12-18|パナソニック株式会社|Passer detection device, passer detection method|
    WO2012002904A1|2010-07-02|2012-01-05|Fotonic I Norden Ab|Device and method for detection of abnormal spatial states of a human body|
    CH703936A2|2010-10-01|2012-04-13|Fastcom Technology Sa|System and method for individualizing people.|
    JP5163971B1|2011-09-12|2013-03-13|オプテックス株式会社|Laser scan sensor|
    JP5263692B2|2011-11-09|2013-08-14|オプテックス株式会社|Laser scan sensor|
    JP5668207B2|2013-01-25|2015-02-12|北陽電機株式会社|Object detection device, distance measuring device, door control device, and automatic door device|
    JP2015213251A|2014-05-02|2015-11-26|株式会社Ihi|Behavior analyzer, monitoring system and amusement system|NL2022442B1|2019-01-24|2020-01-07|Lely Patent Nv|Position determining device|
    CN110889339B|2019-11-12|2020-10-02|南京甄视智能科技有限公司|Head and shoulder detection-based dangerous area grading early warning method and system|
    法律状态:
    2019-02-25| FG| Patent granted|Effective date: 20190129 |
    优先权:
    申请号 | 申请日 | 专利标题
    EP17165849.5|2017-04-10|
    EP17165849.5A|EP3388864A1|2017-04-10|2017-04-10|Method of human body recognition and human body recognition sensor|
    [返回顶部]