Personnel positioning method and device based on overlooking camera, storage medium and electronic equipment

Abstract

The invention provides a personnel positioning method and device based on a top view camera, a storage medium and electronic equipment, and relates to the technical field of indoor positioning, and the method comprises the steps: obtaining a plurality of top views collected by a plurality of cameras for a target region at the same moment, extracting a first feature point corresponding to each to-be-positioned object recognized in the plurality of top views, and obtaining a first feature point corresponding to each to-be-positioned object; obtaining a first feature point set; fusing points meeting a preset matching condition in the first feature point set to obtain a second feature point set comprising a plurality of second feature points; and performing clustering analysis on the second feature points in the second feature point set, and determining position information of the to-be-positioned object in the target area according to a clustering analysis result. According to the personnel positioning method and device based on the overlook camera, the storage medium and the electronic equipment provided by the invention, all personnel in an area range are accurately positioned by analyzing the overlook image collected by the overlook camera, and the position distribution information of the personnel in a relatively large space is accurately provided.

Description

Technical Field

[0001] The present application relates to the field of indoor positioning technology, and in particular to a personnel positioning method, device, storage medium and electronic device based on a bird's-eye view camera. Background Art

[0002] With the development of science and technology, there is a great demand for personnel positioning in shopping malls, scenic spots, hospitals, railway station navigation and smart cities. In related technologies, image-based positioning is usually used to locate people in the monitoring area.

[0003] However, in order to improve the efficiency of personnel detection, the cameras used in the positioning methods in the related technologies are all arranged with oblique viewing angles. And because the camera is a two-dimensional compressed perception of the three-dimensional world, the camera with an oblique viewing angle will lose the original three-dimensional coordinate information in the depth direction. Therefore, in order to obtain the exact two-dimensional plane position of the person, it is necessary to rely on the information of multiple cameras for three-dimensional reconstruction, or use laser radar and other means to supplement the information, and the implementation cost of the solution is relatively high.

[0004] Based on this, there is an urgent need for a personnel positioning method to achieve rapid positioning of personnel at a lower cost. Summary of the invention

[0005] The purpose of this application is to provide a personnel positioning method, device, storage medium and electronic device based on a bird's-eye view camera, which is used to accurately locate all personnel within an area by analyzing the overhead view captured by the overhead camera, and accurately provide position distribution information of personnel in a larger space.

[0006] The present application provides a personnel positioning method based on a top-down camera, comprising: Acquire multiple overhead views of a target area captured by multiple cameras at the same time, and extract first feature points corresponding to each object to be located identified in the multiple overhead views to obtain a first feature point set; one camera captures one overhead view; the points in the first feature point set that meet the preset matching conditions are merged to obtain a second feature point set containing multiple second feature points; cluster analysis is performed on the second feature points in the second feature point set, and the position information of the object to be located in the target area is determined according to the cluster analysis result; wherein the preset matching conditions include: the feature point distance between two first feature points is less than or equal to the preset distance threshold, and the two first feature points are first feature points captured by two cameras whose field of view is adjacent or intersecting; the feature point distance between any two first feature points is a two-dimensional space distance calculated based on the same horizontal plane.

[0007] Optionally, the extracting the first feature points corresponding to each object to be located identified in multiple overhead views to obtain a first feature point set includes: performing target recognition on each overhead view to identify the object to be located in each overhead view; determining the center point of the target recognition frame corresponding to each object to be located as the first feature point corresponding to each object to be located, and generating the first feature point set based on the first feature point corresponding to each object to be located.

[0008] Optionally, the points in the first feature point set that meet the preset matching conditions are fused to obtain a second feature point set containing multiple second feature points, including: calculating the feature point distance between any two first feature points in the first feature point set, and screening out first feature points whose feature point distance is less than the preset distance threshold; fusing first feature points whose feature point distance is less than or equal to the preset distance threshold to obtain multiple second feature points, and generating the second feature point set based on the multiple second feature points.

[0009] Optionally, after acquiring multiple overhead views of the target area captured by multiple cameras at the same time, and extracting the first feature points corresponding to each object to be located identified in the multiple overhead views, the method further includes: screening out third feature points in the first feature point set that do not meet the preset matching conditions, and using the third feature points as position confirmation points corresponding to the object to be located; wherein the position confirmation points are used to analyze and confirm the position information of the object to be located.

[0010] Optionally, cluster analysis is performed on the second feature points in the second feature point set, and the location information of the object to be located in the target area is determined based on the cluster analysis result, including: determining the range of each second feature point based on a preset radius distance, and determining the second feature point whose number of second feature points in the neighborhood is greater than the minimum number of feature points as the first core feature point; and determining the location information of the object to be located based on the feature area formed by the first core feature point and all the second feature points in the neighborhood of the first core feature point.

[0011] Optionally, the determining the location information of the object to be located based on the feature area formed by the first core feature point and all the second feature points within the neighborhood of the first core feature point includes: determining the location information of the object to be located based on the coverage area formed by the first core feature point, all the second feature points within the neighborhood of the first core feature point, and all the second feature points within the neighborhood of the second core feature point.

[0012] Optionally, before fusing the points in the first feature point set that meet the preset matching conditions to obtain a second feature point set containing multiple second feature points, the method also includes: determining the preset distance threshold based on target parameters of the multiple cameras; wherein the target parameters include at least one of the following: camera resolution, camera field of view, camera geometric calibration, and camera imaging delay.

[0013] The present application also provides a personnel positioning device based on a top-down camera, comprising: An acquisition module is used to acquire multiple overhead views of a target area acquired by multiple cameras at the same time; a feature point extraction module is used to extract the first feature points corresponding to each object to be located identified in the multiple overhead views to obtain a first feature point set; one camera acquires one overhead view; a feature point fusion module fuses the points in the first feature point set that meet the preset matching conditions to obtain a second feature point set containing multiple second feature points; a positioning module is used to perform cluster analysis on the second feature points in the second feature point set, and determine the position information of the object to be located in the target area according to the cluster analysis results; wherein the preset matching conditions include: the feature point distance between two first feature points is less than or equal to the preset distance threshold, and the two first feature points are first feature points acquired by two cameras whose field of view is adjacent or intersecting; the feature point distance between any two first feature points is a two-dimensional space distance calculated based on the same horizontal plane.

[0014] Optionally, the feature point extraction module is specifically used to perform target recognition on each overhead view to identify the objects to be located in each overhead view; the feature point extraction module is also specifically used to determine the center point of the target recognition frame corresponding to each object to be located as the first feature point corresponding to each object to be located, and generate the first feature point set based on the first feature point corresponding to each object to be located.

[0015] Optionally, the feature point fusion module is specifically used to calculate the feature point distance between any two first feature points in the first feature point set, and to screen out first feature points whose feature point distance is less than the preset distance threshold; the feature point fusion module is also specifically used to fuse first feature points whose feature point distance is less than or equal to the preset distance threshold to obtain multiple second feature points, and generate the second feature point set based on the multiple second feature points.

[0016] Optionally, the feature point extraction module is also used to screen out a third feature point in the first feature point set that does not meet the preset matching conditions, and use the third feature point as a position confirmation point of the corresponding object to be located; wherein the position confirmation point is used to analyze and confirm the position information of the object to be located.

[0017] Optionally, the positioning module is specifically used to determine the range of each second feature point based on a preset radius distance, and determine the second feature point whose number of second feature points in the neighborhood is greater than the minimum number of feature points as the first core feature point; the positioning module is also specifically used to determine the location information of the object to be located based on the first core feature point and the feature area formed by all second feature points in the neighborhood of the first core feature point.

[0018] Optionally, the positioning module is specifically used to determine the location information of the object to be positioned based on the coverage formed by the first core feature point, all second feature points within the neighborhood of the first core feature point, and all second feature points within the neighborhood of the second core feature point.

[0019] Optionally, the device also includes: a parameter determination module; the parameter determination module is used to determine the preset distance threshold based on target parameters of the multiple cameras; wherein the target parameters include at least one of the following: camera resolution, camera field of view, camera geometric calibration, camera imaging delay.

[0020] The present application also provides a computer program product, including a computer program / instruction, which, when executed by a processor, implements the steps of any of the above-mentioned methods for locating personnel based on a bird's-eye view camera.

[0021] The present application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps of any of the above-mentioned methods for locating personnel based on a bird's-eye view camera are implemented.

[0022] The present application also provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the steps of any of the above-mentioned methods for locating personnel based on a bird's-eye view camera are implemented.

[0023] The personnel positioning method, device, storage medium and electronic device based on the overhead camera provided in the present application first obtain multiple overhead views collected by multiple cameras at the same time for the target area, and extract the first feature points corresponding to each object to be located identified in the multiple overhead views to obtain a first feature point set; one camera collects one overhead view; then, the points in the first feature point set that meet the preset matching conditions are fused to obtain a second feature point set containing multiple second feature points; finally, the second feature points in the second feature point set are clustered and analyzed, and the position information of the object to be located in the target area is determined according to the cluster analysis results; wherein the preset matching conditions include: the feature point distance between the two first feature points is less than or equal to the preset distance threshold, and the two first feature points belong to the first feature points collected by two cameras whose field of view is adjacent or intersecting; the feature point distance between any two first feature points is the two-dimensional space distance calculated based on the same horizontal plane. In this way, by analyzing the overhead view collected by the overhead camera, all personnel within the area can be accurately located, and the position distribution information of personnel in a larger space can be accurately provided. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the technical solutions in the present application or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0025] Figure 1 This is one of the flow charts of the personnel positioning method based on the overhead camera provided in this application; Figure 2 This is the second flow chart of the personnel positioning method based on the overhead camera provided by the present application; Figure 3 It is a structural schematic diagram of a personnel positioning device based on a top-down camera provided in the present application; Figure 4 It is a structural schematic diagram of the electronic device provided by this application. DETAILED DESCRIPTION

[0026] In order to make the purpose, technical solutions and advantages of this application clearer, the technical solutions in this application will be clearly and completely described below in conjunction with the drawings in this application. Obviously, the described embodiments are part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0027] The terms "first", "second", etc. in the specification and claims of the present application are used to distinguish similar objects, and are not used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable under appropriate circumstances, so that the embodiments of the present application can be implemented in an order other than those illustrated or described here, and the objects distinguished by "first", "second", etc. are generally of one type, and the number of objects is not limited. For example, the first object can be one or more. In addition, "and / or" in the specification and claims represents at least one of the connected objects, and the character " / " generally indicates that the objects associated with each other are in an "or" relationship.

[0028] The significant advantage of the image-based personnel positioning system is that it is non-contact and can accurately provide the number of people in the area. Thanks to the rapid progress of machine learning technology, the recognition algorithm using convolutional neural networks can use surveillance videos to achieve a high detection rate for head and shoulder targets. Due to the limited coverage of a single camera, cameras can be placed at major entrances and exits and inside the room, and then data fusion can be performed using Bayesian estimation to improve detection accuracy. Furthermore, by using multiple calibrated cameras with overlapping perspectives, the probability density-based personnel positioning method can provide information on the number and gridded positions of multiple people.

[0029] However, in order to improve the efficiency of personnel detection, the cameras used in the above-mentioned personnel positioning system are all arranged with oblique viewing angles. However, since the camera is a two-dimensional compressed perception of the three-dimensional world, the camera with an oblique viewing angle will lose the original three-dimensional coordinate information in the depth direction. Therefore, in order to obtain the exact three-dimensional position of the personnel, it is necessary to rely on the information of multiple cameras for three-dimensional reconstruction, or use laser radar and other means to supplement the information. It can still be used in small scenes, but it is difficult to apply in large spaces with a large number of people and complex scenes. In addition, since one area needs to be covered by multiple cameras, the system cost is very high and the data processing is extremely difficult. In summary, it is still a difficult problem to locate personnel indoors, especially in larger spaces.

[0030] In response to the above technical problems existing in the related technologies, the embodiments of the present application propose a personnel positioning method based on a bird's-eye view camera. The use of a bird's-eye view camera can avoid the problem of loss of personnel planar position information caused by the side view of the camera, and can accurately provide the position distribution information of personnel in a larger space. The actual operation under various working conditions proves that the system has the characteristics of low cost, high positioning accuracy, and fast response speed, providing a key technical foundation for the efficient creation of demand-oriented indoor environments.

[0031] In the present application embodiment, Figure 1As shown in the figure, multiple cameras with vertical downward viewing angles are arranged on the roof. Since the projection of people in the horizontal direction in the room does not occupy a large area, and in most cases they are standing or sitting. The conventional baffles and partitions in the room are in the vertical direction, and have limited blocking effect on the vertical downward viewing angle. Therefore, the cameras arranged vertically downward can better capture the images of people in the room.

[0032] Since people in the room usually only stay near the ground. When the camera is pointed vertically downward, the projection position of the person is directly related to the horizontal position of the floor. After completing the distortion correction of the camera lens and sensor, it can be almost considered that the position of the person in the camera is proportional to the horizontal position in the actual space. Therefore, after completing the calibration of the camera's external parameters, a direct correspondence between the photo coordinate point and the horizontal position in the actual space can be established. In the figure below, it can be seen more clearly that since the person is in the lower part of the room, the deviation in the horizontal direction of the projection generated by the different positions of the human body captured by the camera is already smaller than the width of the human body, which can meet the requirements of environmental creation control based on the position of the person. Therefore, the method of arranging the camera vertically downward from the top combined with the camera's calibration information can solve the problem of information loss caused by the camera's compressed sensing in the optical axis direction, so that a single camera can complete the positioning of the person in the room in the horizontal direction.

[0033] Considering that the imaging area of ​​a camera is usually rectangular, multiple cameras can be used to achieve wider coverage according to the specific shape of the room and the range that needs to be covered. And because the camera's shooting range and the size of the room can be accurately estimated in advance, the system can be conveniently used in larger scenes. At the same time, because the image is taken from the top, it is not easy to accurately obtain the identity information of the target object, which has a certain privacy protection capability.

[0034] The personnel positioning method based on a bird's-eye view camera provided in the embodiment of the present application is described in detail below through specific embodiments and their application scenarios in conjunction with the accompanying drawings.

[0035] like Figure 2 As shown, an embodiment of the present application provides a method for positioning a person based on a bird's-eye view camera, which may include the following steps 201 to 203: Step 201: Acquire multiple top views of a target area captured by multiple cameras at the same time, and extract first feature points corresponding to each object to be located identified in the multiple top views to obtain a first feature point set.

[0036] Among them, a camera captures a top view.

[0037] Exemplarily, the multiple cameras are cameras arranged at the top of the target area, and each camera is used to collect overhead views of different sub-areas in the target area. There may be overlapping areas between the overhead views collected by different cameras.

[0038] For example, after obtaining the top view captured by each camera at the same time, the first feature point corresponding to each object to be located identified in the multiple top views can be extracted. The multiple top views contain one or more objects to be located, and the objects to be located in the top views can be identified by the target detection model, and each object to be located is extracted according to the identification result.

[0039] Specifically, the step of extracting the first feature point corresponding to each object to be located in the above step 201 may further include the following steps 201a1 and 201a2: Step 201a1: perform target recognition on each top view to identify the object to be located in each top view.

[0040] Step 201a2: determine the center point of the target recognition frame corresponding to each object to be located as the first feature point corresponding to each object to be located, and generate the first feature point set based on the first feature point corresponding to each object to be located.

[0041] Exemplarily, after performing target detection on the above-mentioned multiple top views using the target detection model, a target detection frame corresponding to each object to be located can be generated, and then the center point of the target detection frame of each object to be located is used as the first feature point corresponding to each object to be located.

[0042] It should be noted that, in the embodiment of the present application, the first feature point corresponding to each positioning can be not only the center point of the target detection frame, but also a vertex or any point within the range of the target detection frame.

[0043] Step 202: Fusing the points in the first feature point set that meet a preset matching condition to obtain a second feature point set including a plurality of second feature points.

[0044] Among them, the preset matching conditions include: the feature point distance between two first feature points is less than or equal to a preset distance threshold, and the two first feature points are first feature points collected by two cameras with adjacent or intersecting field of view; the feature point distance between any two first feature points is a two-dimensional space distance calculated based on the same horizontal plane.

[0045] Exemplarily, after obtaining the first feature point corresponding to each object to be located, due to the overlap of the camera's shooting area, there may be a situation where one object to be located corresponds to multiple first feature points. Therefore, it is necessary to perform cluster analysis on the situation where the object to be located corresponds to multiple first feature points, and determine the more accurate location information of each object to be located based on the classification results.

[0046] Exemplarily, in order to distinguish different situations of the object to be located, it is necessary to cyclically calculate the feature point distance between each first feature point and other first feature points, and perform further judgment based on the feature point distance.

[0047] Specifically, the above step 202 may further include the following steps 202a1 and 202a2: Step 202a1, calculating the feature point distance between any two first feature points in the first feature point set, and screening out first feature points whose feature point distance is less than the preset distance threshold.

[0048] Step 202a2: merge the first feature points whose feature point distance is less than or equal to the preset distance threshold to obtain a plurality of second feature points, and generate the second feature point set based on the plurality of second feature points.

[0049] Exemplarily, for any two first feature points whose distance between the feature points is less than a preset distance threshold, a second feature point may be generated based on the two first feature points, that is, a second feature point is generated through two first feature points with a relatively close distance.

[0050] In a possible implementation, a first feature point that is far away from other first feature points can be directly used as a position confirmation point corresponding to an object to be located. It should be noted that if the positions of people are very scattered, so that a person can only be photographed by one camera, there may be only one second feature point in the neighborhood of the first core feature point.

[0051] Exemplarily, after the above step 201, the personnel positioning method based on the overhead camera provided in the embodiment of the present application may further include the following step S1: Step S1: Filter out the third feature point that does not meet the preset matching condition in the first feature point set, and use the third feature point as a position confirmation point of the corresponding object to be located.

[0052] The location confirmation point is used to analyze and confirm the location information of the object to be located.

[0053] Exemplarily, if the distances between a first feature point and other first feature points are relatively far, the first feature point is directly retained as a discrete personnel position confirmation point.

[0054] Step 203: Perform cluster analysis on the second feature points in the second feature point set, and determine the position information of the object to be located in the target area according to the cluster analysis result.

[0055] Exemplarily, after obtaining the second feature points after merging the first feature points, a cluster analysis may be performed on the second feature points, and the object to be located may be located according to the cluster analysis result.

[0056] Specifically, the cluster analysis algorithm in step 203 includes the following steps 203a and 203b: Step 203a: determine the neighborhood range of each second feature point based on a preset radius distance, and determine the second feature points whose number in the neighborhood range is greater than the minimum number of feature points as first core feature points.

[0057] Step 203b: Determine the position information of the object to be located based on the first core feature point and the feature area formed by all second feature points within the neighborhood of the first core feature point.

[0058] Specifically, the above step 203b may further include the following step 203b1: Step 203b1: Determine the position information of the object to be located based on the coverage formed by the first core feature point, all second feature points within the neighborhood of the first core feature point, and all second feature points within the neighborhood of the second core feature point.

[0059] The domain range of the second core feature point includes at least one second feature point in the neighborhood range of the first core feature point.

[0060] It should be noted that in the embodiment of the present application, feature point fusion can be performed on the first core feature point, all second feature points within the neighborhood of the first core feature point, and all second feature points within the neighborhood of the second core feature point, and they can be fused into one feature point, and the feature point can be determined as the position confirmation point of the object to be located.

[0061] Optionally, in this embodiment, before calculating the distances between the first feature points and fusing the first feature points, the influence of the system error needs to be considered.

[0062] For example, before the above step 202, the personnel positioning method based on the overhead camera provided in the embodiment of the present application may further include the following step 205: Step 205: Determine the preset distance threshold based on the target parameters of the multiple cameras.

[0063] The target parameter includes at least one of the following: camera resolution, camera field of view, camera geometric calibration, and camera imaging delay.

[0064] For example, in a multi-camera personnel positioning system, the above-mentioned preset distance threshold is usually related to the following factors: Camera resolution: The higher the resolution, the more details can be captured, which in theory allows for more accurate positioning. Camera field of view (FOV): The wider the field of view, the larger the area a single camera can cover, but some positioning accuracy may be sacrificed. Geometric calibration between cameras: The relative position and posture between cameras need to be accurately calibrated to ensure the accuracy of positioning data. Synchronicity: The images captured by multiple cameras need to be synchronized, and time deviations will affect the accuracy of positioning. Environmental factors: Lighting conditions, obstructions, etc. will affect the quality of images captured by the camera, and thus affect the positioning results.

[0065] In practical applications, the processing power of the algorithm also needs to be considered, how to quickly and effectively process and fuse data from different cameras. For example, some advanced fusion methods use deep learning technology to improve the accuracy and efficiency of fusion. In addition, multi-sensor fusion positioning methods are also a research hotspot, which is not limited to visual sensors, but may also include other types of sensors such as lidar and millimeter-wave radar.

[0066] The personnel positioning method based on the overhead camera provided in the embodiment of the present application first obtains multiple overhead views collected by multiple cameras at the same time for the target area, and extracts the first feature points corresponding to each object to be located identified in the multiple overhead views to obtain a first feature point set; one camera collects one overhead view; then, the points in the first feature point set that meet the preset matching conditions are fused to obtain a second feature point set containing multiple second feature points; finally, cluster analysis is performed on the second feature points in the second feature point set, and the position information of the object to be located in the target area is determined according to the cluster analysis results; wherein the preset matching conditions include: the feature point distance between two first feature points is less than or equal to the preset distance threshold, and the two first feature points belong to the first feature points collected by two cameras whose field of view is adjacent or intersecting; the feature point distance between any two first feature points is the two-dimensional space distance calculated based on the same horizontal plane. In this way, by analyzing the overhead view collected by the overhead camera, all personnel within the area can be accurately located, and the position distribution information of personnel in a larger space can be accurately provided.

[0067] It should be noted that the personnel positioning method based on a bird's-eye view camera provided in the embodiment of the present application can be executed by a personnel positioning device based on a bird's-eye view camera, or a control module in the personnel positioning device based on a bird's-eye view camera for executing the personnel positioning method based on a bird's-eye view camera. In the embodiment of the present application, the personnel positioning device based on a bird's-eye view camera executing the personnel positioning method based on a bird's-eye view camera is taken as an example to illustrate the personnel positioning device based on a bird's-eye view camera provided in the embodiment of the present application.

[0068] It should be noted that, in the embodiments of the present application, the personnel positioning methods based on overhead cameras shown in the above-mentioned method drawings are all illustrated by taking one of the drawings in the embodiments of the present application as an example. In specific implementation, the personnel positioning methods based on overhead cameras shown in the above-mentioned method drawings can also be implemented in combination with any other drawings that can be combined as shown in the above-mentioned embodiments, which will not be repeated here.

[0069] The following is a description of the personnel positioning device based on a bird's-eye view camera provided in the present application. The personnel positioning method based on a bird's-eye view camera described below and described above can refer to each other.

[0070] Figure 3 A schematic diagram of the structure of a personnel positioning device based on a top-down camera provided in an embodiment of the present application, such as Figure 3 As shown, specifically including: The acquisition module 301 is used to acquire multiple overhead views collected by multiple cameras at the same time for the target area; the feature point extraction module 302 is used to extract the first feature points corresponding to each object to be located identified in the multiple overhead views to obtain a first feature point set; one camera collects one overhead view; the feature point fusion module 303 fuses the points in the first feature point set that meet the preset matching conditions to obtain a second feature point set containing multiple second feature points; the positioning module 304 is used to perform cluster analysis on the second feature points in the second feature point set, and determine the position information of the object to be located in the target area according to the cluster analysis results; wherein the preset matching conditions include: the feature point distance between two first feature points is less than or equal to the preset distance threshold, and the two first feature points are first feature points collected by two cameras whose field of view is adjacent or intersecting; the feature point distance between any two first feature points is a two-dimensional space distance calculated based on the same horizontal plane.

[0071] Optionally, the feature point extraction module 302 is specifically used to perform target recognition on each overhead view to identify the objects to be located in each overhead view; the feature point extraction module 302 is also specifically used to determine the center point of the target recognition frame corresponding to each object to be located as the first feature point corresponding to each object to be located, and generate the first feature point set based on the first feature point corresponding to each object to be located.

[0072] Optionally, the feature point fusion module 303 is specifically used to calculate the feature point distance between any two first feature points in the first feature point set, and to screen out first feature points whose feature point distance is less than the preset distance threshold; the feature point fusion module 303 is also specifically used to fuse first feature points whose feature point distance is less than or equal to the preset distance threshold to obtain multiple second feature points, and generate the second feature point set based on the multiple second feature points.

[0073] Optionally, the feature point extraction module 302 is also used to screen out a third feature point in the first feature point set that does not meet the preset matching condition, and use the third feature point as a position confirmation point of the corresponding object to be located; wherein the position confirmation point is used to analyze and confirm the position information of the object to be located.

[0074] Optionally, the positioning module 304 is specifically used to determine the range of each second feature point based on a preset radius distance, and determine the second feature point whose number of second feature points in the neighborhood is greater than the minimum number of feature points as the first core feature point; the positioning module 304 is also specifically used to determine the location information of the object to be located based on the first core feature point and the feature area formed by all second feature points in the neighborhood of the first core feature point.

[0075] Optionally, the positioning module 304 is specifically used to determine the location information of the object to be positioned based on the coverage formed by the first core feature point, all second feature points within the neighborhood of the first core feature point, and all second feature points within the neighborhood of the second core feature point.

[0076] Optionally, the device also includes: a parameter determination module; the parameter determination module is used to determine the preset distance threshold based on target parameters of the multiple cameras; wherein the target parameters include at least one of the following: camera resolution, camera field of view, camera geometric calibration, camera imaging delay.

[0077] The personnel positioning device based on the overhead camera provided in the present application first obtains multiple overhead views collected by multiple cameras at the same time for the target area, and extracts the first feature points corresponding to each object to be located identified in the multiple overhead views to obtain a first feature point set; one camera collects one overhead view; then, the points in the first feature point set that meet the preset matching conditions are fused to obtain a second feature point set containing multiple second feature points; finally, the second feature points in the second feature point set are clustered and analyzed, and the position information of the object to be located in the target area is determined according to the cluster analysis results; wherein the preset matching conditions include: the feature point distance between two first feature points is less than or equal to the preset distance threshold, and the two first feature points belong to the first feature points collected by two cameras whose field of view is adjacent or intersecting; the feature point distance between any two first feature points is the two-dimensional space distance calculated based on the same horizontal plane. In this way, by analyzing the overhead view collected by the overhead camera, all personnel within the area can be accurately located, and the position distribution information of personnel in a larger space can be accurately provided.

[0078] Figure 4 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 4 As shown, the electronic device may include: a processor (processor) 410 , a communication interface (Communications Interface) 420 , a memory (memory) 430 and a communication bus 440 , wherein the processor 410 , the communication interface 420 , and the memory 430 communicate with each other through the communication bus 440 . The processor 410 can call the logic instructions in the memory 430 to execute a personnel positioning method based on a bird's-eye view camera, the method comprising: obtaining multiple bird's-eye views of a target area captured by multiple cameras at the same time, and extracting the first feature points corresponding to each object to be positioned identified in the multiple bird's-eye views to obtain a first feature point set; one camera captures one bird's-eye view; the points in the first feature point set that meet the preset matching conditions are merged to obtain a second feature point set containing multiple second feature points; cluster analysis is performed on the second feature points in the second feature point set, and the position information of the object to be positioned in the target area is determined according to the cluster analysis results; wherein the preset matching conditions include: the feature point distance between two first feature points is less than or equal to a preset distance threshold, and the two first feature points are first feature points captured by two cameras whose field of view is adjacent or intersecting; the feature point distance between any two first feature points is a two-dimensional space distance calculated based on the same horizontal plane.

[0079] In addition, the logic instructions in the above-mentioned memory 430 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0080] On the other hand, the present application also provides a computer program product, which includes a computer program stored on a computer-readable storage medium, and the computer program includes program instructions. When the program instructions are executed by a computer, the computer can execute the personnel positioning method based on the overhead camera provided by the above methods, the method including: obtaining multiple overhead views collected by multiple cameras for the target area at the same time, and extracting the first feature points corresponding to each object to be located identified in the multiple overhead views to obtain a first feature point set; one camera collects one overhead view; after fusing the points in the first feature point set that meet the preset matching conditions, a second feature point set containing multiple second feature points is obtained; cluster analysis is performed on the second feature points in the second feature point set, and the position information of the object to be located in the target area is determined according to the cluster analysis results; wherein the preset matching conditions include: the feature point distance between two first feature points is less than or equal to the preset distance threshold, and the two first feature points belong to the first feature points collected by two cameras whose field of view is adjacent or intersecting; the feature point distance between any two first feature points is a two-dimensional space distance calculated based on the same horizontal plane.

[0081] On the other hand, the present application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the above-mentioned personnel positioning method based on a bird's-eye view camera, the method comprising: obtaining multiple bird's-eye views captured by multiple cameras at the same time for a target area, and extracting first feature points corresponding to each object to be positioned identified in the multiple bird's-eye views to obtain a first feature point set; one camera captures one bird's-eye view; points in the first feature point set that meet preset matching conditions are merged to obtain a second feature point set containing multiple second feature points; cluster analysis is performed on the second feature points in the second feature point set, and the position information of the object to be positioned in the target area is determined based on the cluster analysis results; wherein the preset matching conditions include: the feature point distance between two first feature points is less than or equal to a preset distance threshold, and the two first feature points are first feature points captured by two cameras whose field of view is adjacent or intersecting; the feature point distance between any two first feature points is a two-dimensional space distance calculated based on the same horizontal plane.

[0082] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0083] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0084] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A personnel positioning method based on a bird's-eye view camera, characterized in that: include: Acquire multiple top views of the target area collected by multiple cameras at the same time, and extract first feature points corresponding to each object to be located identified in the multiple top views to obtain a first feature point set; One camera captures a top-down view; Merging the points in the first feature point set that meet the preset matching condition to obtain a second feature point set including a plurality of second feature points; Performing cluster analysis on the second feature points in the second feature point set, and determining position information of the object to be located in the target area according to the cluster analysis result; Among them, the preset matching conditions include: the feature point distance between two first feature points is less than or equal to a preset distance threshold, and the two first feature points are first feature points collected by two cameras with adjacent or intersecting field of view; the feature point distance between any two first feature points is a two-dimensional space distance calculated based on the same horizontal plane.

2. The method according to claim 1, characterized in that The step of extracting first feature points corresponding to each object to be located identified in the plurality of top views to obtain a first feature point set includes: Perform target recognition on each top view to identify the object to be located in each top view; The center point of the target recognition frame corresponding to each object to be located is determined as the first feature point corresponding to each object to be located, and the first feature point set is generated based on the first feature point corresponding to each object to be located.

3. The method according to claim 1, characterized in that The step of fusing the points in the first feature point set that meet the preset matching condition to obtain a second feature point set including a plurality of second feature points comprises: Calculating a feature point distance between any two first feature points in the first feature point set, and screening out first feature points whose feature point distance is less than the preset distance threshold; The first feature points whose feature point distance is less than or equal to the preset distance threshold are merged to obtain a plurality of second feature points, and the second feature point set is generated based on the plurality of second feature points.

4. The method according to claim 1, characterized in that: After acquiring a plurality of top views collected by a plurality of cameras at the same time for the target area and extracting a first feature point corresponding to each object to be located identified in the plurality of top views, the method further includes: Filtering out a third feature point that does not meet the preset matching condition in the first feature point set, and using the third feature point as a position confirmation point of the corresponding object to be located; The location confirmation point is used to analyze and confirm the location information of the object to be located.

5. The method according to claim 1, characterized in that The performing cluster analysis on the second feature points in the second feature point set, and determining the position information of the object to be located in the target area according to the cluster analysis result, includes: Determine the area range of each second feature point based on a preset radius distance, and determine the second feature points whose number of second feature points in the neighborhood range is greater than the minimum number of feature points as first core feature points; Based on the first core feature point and a feature area formed by all second feature points within a neighborhood of the first core feature point, position information of the object to be located is determined.

6. The method according to claim 5, characterized in that The determining the position information of the object to be located based on the first core feature point and the feature area formed by all second feature points within the neighborhood of the first core feature point includes: The position information of the object to be located is determined based on a coverage area formed by the first core feature point, all second feature points within a neighborhood of the first core feature point, and all second feature points within a neighborhood of the second core feature point.

7. The method according to claim 1, characterized in that Before fusing the points in the first feature point set that meet the preset matching condition to obtain a second feature point set including a plurality of second feature points, the method further includes: Determining the preset distance threshold based on target parameters of the multiple cameras; The target parameter includes at least one of the following: camera resolution, camera field of view, camera geometric calibration, and camera imaging delay.

8. A personnel positioning device based on a top-down camera, characterized in that: The device comprises: An acquisition module is used to acquire multiple top views of a target area collected by multiple cameras at the same time; A feature point extraction module is used to extract the first feature points corresponding to each object to be located identified in multiple top views to obtain a first feature point set; one camera captures one top view; A feature point fusion module, which fuses the points in the first feature point set that meet the preset matching condition to obtain a second feature point set including a plurality of second feature points; A positioning module, used for performing cluster analysis on the second feature points in the second feature point set, and determining the position information of the object to be positioned in the target area according to the cluster analysis result; Among them, the preset matching conditions include: the feature point distance between two first feature points is less than or equal to a preset distance threshold, and the two first feature points are first feature points collected by two cameras with adjacent or intersecting field of view; the feature point distance between any two first feature points is a two-dimensional space distance calculated based on the same horizontal plane.

9. An electronic device, characterized in that: The method comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps of the personnel positioning method based on a bird's-eye view camera as claimed in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium, characterized in that: A computer program is stored thereon, and when the computer program is executed by a processor, the steps of the personnel positioning method based on a bird's-eye view camera as claimed in any one of claims 1 to 7 are implemented.

Images