Real-time target tracking method, device and equipment

Abstract

The embodiment of the present application provides a method, device and equipment for real-time target tracking, the method comprising: an electronic device obtains historical trajectory data corresponding to multiple cameras, and determines the target topological relationship corresponding to the multiple cameras based on the historical trajectory data; determines the target to be tracked in the first camera based on the target feature information, and determines the feature information and position information corresponding to the target to be tracked; determines the target-associated camera corresponding to the first camera based on the position information and the target topological relationship; in the target-associated camera, determines the real-time position information of the target to be tracked based on the feature information. In this way, the electronic device can automatically determine the target topological relationship between multiple cameras based on the historical trajectory data of the target, without the need for users to manually repeat labeling, which can save human resources, and at the same time can improve the real-time performance of target tracking, and the scope of application of scenarios is also wider.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a method, device and equipment for real-time target tracking. Background Art

[0002] In related technologies, in order to facilitate the rapid tracking of specific personnel, staff usually mark the positional relationship between each camera through manual labeling. This manual labeling (i.e., manual labeling) method consumes a lot of human resources, and because the cameras in offline scenarios are often adjusted and modified, staff need to repeat manual labeling, resulting in low real-time tracking of specific personnel and a small scope of application. Summary of the invention

[0003] Multiple aspects of the present application provide a real-time target tracking method, device and equipment to realize automatic determination of camera topological relationships without manual marking, thus saving human resources and improving the real-time performance and scene applicability of target tracking.

[0004] In a first aspect, an embodiment of the present application provides a method for real-time tracking of a target, comprising:

[0005] Acquire historical trajectory data corresponding to a plurality of cameras, and determine a target topological relationship corresponding to the plurality of cameras according to the historical trajectory data;

[0006] Determining a target to be tracked based on the target feature information in the first camera, and determining feature information and position information corresponding to the target to be tracked;

[0007] Based on the location information and the target topological relationship, determining a target-associated camera corresponding to the first camera;

[0008] In the target-associated camera, the real-time position information of the target to be tracked is determined based on the feature information.

[0009] In a possible implementation, the acquiring of historical trajectory data corresponding to the plurality of cameras includes:

[0010] Acquire the track point information reported by each camera; the track point information includes the coordinate information and time information of each track point;

[0011] The track point information within a preset historical period is obtained to obtain the historical track data.

[0012] In a possible implementation, determining the target topological relationship corresponding to the multiple cameras according to the historical trajectory data includes:

[0013] Merging the historical trajectory data corresponding to the multiple cameras to obtain trajectory session data corresponding to each target;

[0014] For each camera, based on the historical trajectory data of the camera, determining the effective shooting area corresponding to the camera;

[0015] According to the effective shooting area and the trajectory session data, the association relationship between the effective shooting area of ​​the camera and other cameras is determined, and the target topological relationship corresponding to the multiple cameras is obtained.

[0016] In a possible implementation, determining the effective shooting area corresponding to the camera based on the historical trajectory data of the camera includes:

[0017] Dividing the overall shooting area of ​​the camera into multiple sub-shooting areas;

[0018] Among the multiple sub-shooting areas, deleting the sub-shooting area that does not include the target trajectory point, to obtain the target sub-shooting area;

[0019] The target sub-shooting areas are merged to obtain an effective shooting area corresponding to the camera.

[0020] In a possible implementation, determining the association relationship between the effective shooting area of ​​the camera and other cameras according to the effective shooting area and the trajectory session data, and obtaining the target topological relationship corresponding to the multiple cameras includes:

[0021] In the effective shooting area of ​​the current camera, determine the target effective shooting area corresponding to the last track point of each track session data;

[0022] Determine, according to the trajectory conversation data, the associated cameras corresponding to the effective shooting area of ​​the target, and determine the trajectory cross-lens time and the number of trajectories corresponding to each of the associated cameras;

[0023] Determine the target association probability between the target effective shooting area and the associated camera based on the trajectory cross-lens time and the number of trajectories;

[0024] According to the target association probability, the target topological relationship corresponding to the multiple cameras is determined.

[0025] In a possible implementation, determining the target to be tracked based on the target feature information in the first camera includes:

[0026] Collecting target feature information of each target in the first camera;

[0027] In the case where the target characteristic information matches the target characteristic information corresponding to the target person in the preset database, the target to which the target characteristic information belongs is determined as the target to be tracked.

[0028] In a possible implementation manner, determining the target-associated camera corresponding to the first camera based on the location information and the target topological relationship includes:

[0029] Determine, according to the position information, a target effective shooting area corresponding to the target to be tracked in the first camera;

[0030] Based on the target topological relationship and the target effective shooting area, a target-associated camera corresponding to the target to be tracked is determined.

[0031] In a possible implementation manner, determining the real-time position information of the target to be tracked based on the feature information in the target-associated camera includes:

[0032] Determine the target association probability corresponding to each target-associated camera according to the target topological relationship;

[0033] Searching for the feature information in each target-associated camera in descending order of target-associated probability;

[0034] In the case that target feature information exists in the target-associated camera and matches feature information of the target to be tracked, the real-time position information of the target to be tracked is determined.

[0035] In a second aspect, an embodiment of the present application provides a target real-time tracking device, comprising:

[0036] An acquisition module, used to acquire historical trajectory data corresponding to a plurality of cameras, and determine a target topological relationship corresponding to the plurality of cameras according to the historical trajectory data;

[0037] A first determination module, configured to determine a target to be tracked based on target feature information in a first camera, and determine feature information and position information corresponding to the target to be tracked;

[0038] A second determination module, configured to determine a target-associated camera corresponding to the first camera based on the location information and the target topological relationship;

[0039] The third determination module is used to determine the real-time position information of the target to be tracked based on the feature information in the target-associated camera.

[0040] In a possible implementation manner, the acquisition module is specifically configured to:

[0041] Acquire the track point information reported by each camera; the track point information includes the coordinate information and time information of each track point;

[0042] The track point information within a preset historical period is obtained to obtain the historical track data.

[0043] In a possible implementation manner, the acquisition module is specifically configured to:

[0044] Merging the historical trajectory data corresponding to the multiple cameras to obtain trajectory session data corresponding to each target;

[0045] For each camera, based on the historical trajectory data of the camera, determining the effective shooting area corresponding to the camera;

[0046] According to the effective shooting area and the trajectory session data, the association relationship between the effective shooting area of ​​the camera and other cameras is determined, and the target topological relationship corresponding to the multiple cameras is obtained.

[0047] In a possible implementation manner, the acquisition module is specifically configured to:

[0048] Dividing the overall shooting area of ​​the camera into multiple sub-shooting areas;

[0049] Among the multiple sub-shooting areas, deleting the sub-shooting area that does not include the target trajectory point, to obtain the target sub-shooting area;

[0050] The target sub-shooting areas are merged to obtain an effective shooting area corresponding to the camera.

[0051] In a possible implementation manner, the acquisition module is specifically configured to:

[0052] In the effective shooting area of ​​the current camera, determine the target effective shooting area corresponding to the last track point of each track session data;

[0053] Determine, according to the trajectory conversation data, the associated cameras corresponding to the effective shooting area of ​​the target, and determine the trajectory cross-lens time and the number of trajectories corresponding to each of the associated cameras;

[0054] Determine the target association probability between the target effective shooting area and the associated camera based on the trajectory cross-lens time and the number of trajectories;

[0055] According to the target association probability, the target topological relationship corresponding to the multiple cameras is determined.

[0056] In a possible implementation manner, the first determining module is specifically configured to:

[0057] Collecting target feature information of each target in the first camera;

[0058] In the case where the target characteristic information matches the target characteristic information corresponding to the target person in the preset database, the target to which the target characteristic information belongs is determined as the target to be tracked.

[0059] In a possible implementation manner, the second determining module is specifically configured to:

[0060] Determine, according to the position information, a target effective shooting area corresponding to the target to be tracked in the first camera;

[0061] Based on the target topological relationship and the target effective shooting area, a target-associated camera corresponding to the target to be tracked is determined.

[0062] In a possible implementation manner, the third determining module is specifically configured to:

[0063] Determine the target association probability corresponding to each target-associated camera according to the target topological relationship;

[0064] Searching for the feature information in each target-associated camera in descending order of target-associated probability;

[0065] In the case that target feature information exists in the target-associated camera and matches feature information of the target to be tracked, the real-time position information of the target to be tracked is determined.

[0066] In a third aspect, an embodiment of the present application provides an electronic device, including: a memory and a processor;

[0067] The memory stores computer-executable instructions;

[0068] The processor executes the computer-executable instructions stored in the memory, so that the processor executes the target real-time tracking method described in any one of the first aspects.

[0069] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer execution instructions, which, when executed by a processor, are used to implement the target real-time tracking method described in any one of the first aspects.

[0070] In a fifth aspect, an embodiment of the present application provides a computer program product, including a computer program, which, when executed by a processor, implements the real-time target tracking method shown in any one of the first aspects.

[0071] In an embodiment of the present application, historical trajectory data corresponding to multiple cameras are obtained, and the target topological relationship corresponding to the multiple cameras is determined based on the historical trajectory data; the target to be tracked is determined in the first camera based on the target feature information, and the feature information and position information corresponding to the target to be tracked are determined; based on the position information and the target topological relationship, the target-associated camera corresponding to the first camera is determined; in the target-associated camera, the real-time position information of the target to be tracked is determined based on the feature information. In the present application, the electronic device determines the target topological relationship between the multiple cameras based on the historical trajectory data collected by the multiple cameras, and then determines the target-associated camera that the target to be tracked may reach after the first camera based on the position information of the target to be tracked and the predetermined target topological relationship for the target to be tracked, and then matches and searches based on the feature information of the target to be tracked in the target-associated camera to achieve rapid tracking of the target to be tracked. In this way, the electronic device can automatically determine the target topological relationship between the multiple cameras based on the historical trajectory data of the target, without the need for the user to manually repeat the marking, which can save human resources, and at the same time can improve the real-time performance of target tracking, and the scope of application of the scene is also wider. BRIEF DESCRIPTION OF THE DRAWINGS

[0072] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0073] Figure 1 A schematic diagram of an application scenario provided for an exemplary embodiment of the present application;

[0074] Figure 2 A flowchart of a method for real-time target tracking provided by an exemplary embodiment of the present application;

[0075] Figure 3 A flowchart of another target real-time tracking method provided for an exemplary embodiment of the present application;

[0076] Figure 4 A schematic diagram of an effective shooting area of ​​a camera provided by an exemplary embodiment of the present application;

[0077] Figure 5 A trajectory statistics diagram of a cross-mirror relationship of trajectory conversation data provided by an exemplary embodiment of the present application;

[0078] Figure 6 A schematic diagram of real-time tracking provided for an exemplary embodiment of the present application;

[0079] Figure 7 A schematic diagram of a technical link for real-time target tracking provided by an exemplary embodiment of the present application;

[0080] Figure 8 A schematic diagram of the structure of a target real-time tracking device provided for an exemplary embodiment of the present application;

[0081] Fig. 9 A schematic structural diagram of an electronic device provided for an exemplary embodiment of the present application. DETAILED DESCRIPTION

[0082] In order to make the purpose, technical solutions and advantages of the present application clearer, the technical solutions of the present application will be clearly and completely described below in conjunction with the specific embodiments of the present application and the corresponding drawings. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of this application. It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data need to comply with relevant laws, regulations and standards, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0083] In offline new retail scenarios, after specific personnel enter the supermarket, staff can remotely track them through cameras to avoid the risk of theft. In order to achieve non-perceptible (no disturbance) and rapid tracking and positioning of specific personnel, staff have a clear perception of the distribution of various cameras in offline scenarios such as supermarkets. In related technologies, in order to clarify the distribution of each camera, staff usually calibrate each camera one by one through manual marking and on-site observation, and then perform real-time tracking based on the calibrated multiple cameras. This real-time tracking method of manual marking consumes a lot of human resources, and because the cameras in offline scenarios are often adjusted and modified, staff are required to manually re-mark, resulting in low real-time tracking of specific personnel and a small scope of application in scenarios.

[0084] In order to solve the above problems, the embodiments of the present application provide a method, device and equipment for real-time target tracking. The electronic device determines the target topological relationship between multiple cameras based on the historical trajectory data collected by multiple cameras. Then, for the target to be tracked, based on the location information of the target to be tracked and the target topological relationship, it determines the target-associated camera to which the target to be tracked may move after the first camera. Subsequently, matching and searching are performed in the target-associated camera based on the feature information of the target to be tracked, so as to achieve rapid tracking of the target to be tracked. In this way, the electronic device can automatically determine the target topological relationship between multiple cameras based on the historical trajectory data of the target, without the need for users to manually repeat labeling, which can save human resources, and at the same time can improve the real-time performance of target tracking, and the scope of application of scenarios is also wider.

[0085] Figure 1 A schematic diagram of an application scenario provided by an exemplary embodiment of the present application. Figure 1 As shown, the real-time target tracking method in the related art requires users (such as staff in offline retail scenarios, etc.) to manually mark the location distribution of multiple cameras, and then perform real-time tracking based on the manually marked multiple cameras. This tracking method based on manual marking consumes a lot of human resources, and the real-time performance of target tracking is not strong, and the scene adaptability is not high.

[0086] In the embodiment of the present application, the electronic device automatically determines the target topological relationship between multiple cameras based on the historical trajectory data of multiple cameras. When tracking the target to be tracked, the electronic device can determine the target-associated camera to which the target to be tracked may move after the first camera based on the predetermined target topological relationship and the position information of the target to be tracked. Subsequently, searching and tracking are performed in the target-associated cameras, which can achieve accurate and real-time tracking of the target to be tracked without the need for manual labeling by the user, saving human resources, improving the real-time performance of target tracking, and being applicable to a variety of offline retail tracking scenarios.

[0087] The technical solution shown in the present application is described in detail below through specific embodiments. It should be noted that the following embodiments can exist independently or in combination with each other, and the same or similar contents will not be described repeatedly in different embodiments.

[0088] Figure 2 A flowchart of a method for real-time target tracking provided by an exemplary embodiment of the present application. Figure 2 , the target real-time tracking method may include:

[0089] S201, obtaining historical trajectory data corresponding to multiple cameras, and determining a target topological relationship corresponding to the multiple cameras according to the historical trajectory data.

[0090] The execution subject of the embodiment of the present application may be an electronic device, or a target real-time tracking device set in the electronic device. The target real-time tracking device may be implemented by software, or by a combination of software and hardware. For ease of understanding, hereinafter, the execution subject is an electronic device as an example for explanation, and the electronic device may specifically refer to a server, a mobile phone, etc. The embodiment of the present application does not limit the specific type of the electronic device.

[0091] In an embodiment of the present application, multiple cameras may refer to cameras deployed and installed in different locations of the same supermarket in an offline retail scenario. The multiple cameras are communicatively connected to the electronic device, and each camera can report video streams and data such as trajectory points of each target to the electronic device. Historical trajectory data may refer to target trajectory data collected by each camera within a historical period, and the historical period may refer to the past one day, three days, or seven days, etc. The target topological relationship may refer to the association relationship between each camera, which can characterize the cross-lens relationship and cross-lens movement probability of the target between each camera. In this step, the electronic device can obtain the historical trajectory data of multiple cameras, and the historical trajectory data may be a set of trajectory points corresponding to multiple targets; then the electronic device can merge, analyze and count the historical trajectory data, determine the connectivity relationship between multiple cameras, and obtain the target topological relationship between multiple cameras.

[0092] Specifically, when determining the target topological relationship, the electronic device can first analyze within a single camera, and determine the effective shooting area corresponding to the single camera based on the trajectory points in the historical trajectory data; then the electronic device can merge the historical trajectory data, and merge the trajectory points of the same target into the same trajectory session data; then the electronic device determines the effective shooting area where the last trajectory point corresponding to each target is located in the single camera (i.e., the effective shooting area corresponding to when the trajectory disappears), and can determine the next associated camera that the target reaches from the effective shooting area based on the trajectory session data, so that the trajectory session data of multiple targets are analyzed and counted, and the associated cameras corresponding to each effective shooting area can be obtained, and the target association probability from the effective shooting area to the associated camera can be obtained based on the number of trajectories corresponding to each associated camera, the cross-mirror time consumption and other data, and then the electronic device can count the target association probability between different cameras to obtain the target topological relationship. Of course, the electronic device can also use other methods to determine the target topological relationship, and the specific implementation process can be flexibly set based on actual needs, and the embodiments of the present application are not limited to this.

[0093] S202: Determine a target to be tracked based on the target feature information in the first camera, and determine feature information and position information corresponding to the target to be tracked.

[0094] In the embodiment of the present application, the first camera may refer to the entrance camera of a supermarket in an offline retail scene, or may refer to the camera where the target to be tracked appears for the first time, or may refer to any camera in a shopping mall, and the embodiment of the present application does not limit this. The target to be tracked may refer to a specific person, such as a person who has a specific behavior (such as theft, false scanning, abnormal purchase or missed scanning, etc.) in a historical time. Target feature information may refer to the feature information of each target collected by the camera. The feature information of the target to be tracked may refer to the feature information of a specific person. Among them, the feature information may specifically include multiple features of a person at multiple angles, such as shoe features, body features, head features, and accessory features. The position information of the target to be tracked may refer to the position information of a specific person in the first camera, and may specifically refer to the position of the foot point of the target to be tracked at a certain moment in the video image, which may be specifically represented by image coordinates in an image coordinate system, etc. The embodiment of the present application does not limit the specific representation form of the position information.

[0095] In this step, after determining the target topological relationship between multiple cameras, the electronic device can track the target in real time based on the topological relationship. Specifically, the electronic device can first determine the target to be tracked that needs to be tracked in real time in the video stream of the first camera (for example, in the entrance camera), and can specifically perform identification and matching based on a preset specific personnel database, etc. At the same time, the electronic device can determine the feature information and position information of the target to be tracked in the first camera, and save the feature information and position information to the tracking session. The electronic device can update the position information and feature information in real time as the target to be tracked moves, and quickly track the target to be tracked based on the feature information and position information when the target to be tracked moves across the mirror. The specific determination process of the feature information and position information can refer to identification and detection based on image recognition algorithms or deep learning algorithms, etc., which is not limited in the embodiments of the present application.

[0096] S203: Determine a target-associated camera corresponding to the first camera based on the location information and the target topological relationship.

[0097] In an embodiment of the present application, the target-associated camera may refer to each camera associated with the first camera, and specifically may refer to the next camera that may be reached after the target to be tracked disappears from the first camera. The number of target-associated cameras may be one, two or more, and the embodiment of the present application is not limited to this.

[0098] In this step, after determining the position information and feature information of the target to be tracked in the first camera, the electronic device can determine the target-associated camera that the target to be tracked may reach based on the position information and the target topological relationship. Specifically, the electronic device can determine the effective shooting area of ​​the target to be tracked in the first camera based on the position information, and then determine the target-associated camera corresponding to the first camera based on the association relationship between the effective shooting area and the associated camera in the target topological relationship. Alternatively, the electronic device can determine the target-associated camera corresponding to the first camera based on the association relationship between the first camera and other cameras in the target topological relationship, which is not limited in the embodiments of the present application. In this way, based on the position information of the target to be tracked and the target topological relationship between the predetermined multiple cameras, the electronic device can quickly determine the next target-associated camera that the target to be tracked may reach when the target to be tracked moves across the mirror, and can subsequently search and match in the target-associated camera, which can improve the real-time and accuracy of target tracking.

[0099] Exemplarily, the first camera A may include an effective shooting area A1, an effective shooting area A2 and an effective shooting area A3. In the target topological relationship, the associated cameras corresponding to the effective shooting area A1 in the first camera A may be camera B and camera C, the associated cameras corresponding to the effective shooting area A2 in the first camera A may be camera D, camera E and camera F, and the associated cameras corresponding to the effective shooting area A3 in the first camera A may be camera G, camera H and camera I; on this basis, the electronic device determines, based on the position information of the target to be tracked, that the effective shooting area to which the position information belongs is the effective shooting area A2, so that the electronic device can determine, based on the target topological relationship, that the target-associated cameras that the target to be tracked may reach after the first camera are camera D, camera E and camera F.

[0100] S204: In the target-associated camera, determine the real-time position information of the target to be tracked based on the feature information.

[0101] In an embodiment of the present application, the real-time location information may refer to the location information of the target to be tracked in the target-associated camera. After the electronic device determines each target-associated camera after the first camera of the target to be tracked, it may search and match the feature information of the target to be tracked in the video stream captured by each target-associated camera to determine the real-time location information of the target to be tracked. Based on the processing performance of the electronic device, the electronic device may perform a matching search for the feature information in the order of the target association probability corresponding to each target-associated camera in the target topological relationship from high to low, or may perform a matching search for the feature information in each target-associated camera at the same time, which is not limited in the embodiment of the present application.

[0102] In an embodiment of the present application, an electronic device obtains historical trajectory data corresponding to multiple cameras, and determines the target topological relationship corresponding to the multiple cameras based on the historical trajectory data; determines the target to be tracked in the first camera based on the target feature information, and determines the feature information and position information corresponding to the target to be tracked; determines the target-associated camera corresponding to the first camera based on the position information and the target topological relationship; and determines the real-time position information of the target to be tracked in the target-associated camera based on the feature information. In the present application, the electronic device determines the target topological relationship between the multiple cameras based on the historical trajectory data collected by the multiple cameras, and then determines the target-associated camera that the target to be tracked may reach after the first camera based on the position information and the target topological relationship of the target to be tracked. Subsequently, matching and searching are performed in the target-associated camera based on the feature information of the target to be tracked, so as to realize real-time tracking of the target to be tracked. In this way, the electronic device can automatically determine the target topological relationship between the multiple cameras based on the historical trajectory data of the target, without the need for the user to manually repeat the marking, which can save human resources, and at the same time can improve the real-time performance of target tracking, and the scope of application of scenes is also wider.

[0103] Based on the above embodiments, Figure 3 A flowchart of another target real-time tracking method provided by an exemplary embodiment of the present application. Figure 3 , the target real-time tracking method may include:

[0104] S301, obtaining the track point information reported by each camera; the track point information includes the coordinate information and time information of each track point; obtaining the track point information within a preset historical period to obtain historical track data.

[0105] In the embodiment of the present application, the track point information may refer to the track point data of the target at different times reported by each camera. The track point information may include the coordinate information and time information of each track point, wherein the coordinate information may refer to the image coordinates of the target foot point in the track point, and the time information may refer to the time corresponding to the track point. The preset historical period may refer to the sampling period of the preset track data, which may specifically refer to one day, three days or seven days, etc., which may be flexibly set based on actual needs, and the embodiment of the present application does not limit this.

[0106] In this step, the electronic device can obtain the track point information reported by each camera in the supermarket at each moment, and the track point information includes the coordinate information and time information of the target in the camera in the video stream; afterwards, when it is necessary to determine the target topological relationship between multiple cameras, the electronic device can obtain the track point information within a preset historical period, for example, the electronic device can obtain the track point information within the past day, obtain the historical track data, and then determine the target topological relationship based on the historical track data. In this way, the electronic device generates historical track data by collecting the track point information within the preset historical period, which can ensure the comprehensiveness and reliability of the target track data sample and improve the rationality of the subsequent determination of the target topological relationship.

[0107] S302: Merge the historical trajectory data corresponding to the multiple cameras to obtain the trajectory session data corresponding to each target.

[0108] In the embodiment of the present application, the trajectory session data may refer to the trajectory data of the same target between multiple cameras within a preset historical period. After the electronic device obtains the historical trajectory data of multiple targets, it may first merge the historical trajectory data to determine the trajectory session data of the same target. Specifically, the electronic device may merge the trajectory point information continuously captured by the same target in a single camera to obtain the trajectory segments of the same target in a single camera. After that, for the trajectory segments of each target in multiple cameras, the electronic device may aggregate the trajectory segments of the same target to obtain the trajectory session data corresponding to each target. In the specific merging process, the electronic device may perform merging based on algorithms such as person re-identification (REID) and feature information of the target to achieve rapid determination of the trajectory session data. Of course, the electronic device may also use other algorithms to merge the historical trajectory data, which is not limited in the embodiment of the present application.

[0109] S303: For each camera, determine the effective shooting area corresponding to the camera based on the historical trajectory data of the camera.

[0110] In the embodiment of the present application, the effective shooting area may refer to the shooting area including the target trajectory point in a single camera. In the visual range of the camera, that is, the range that the camera can capture, the camera can capture all images within the range, which may include non-ground areas such as walls and shelves. These non-ground areas are inaccessible to the target foot points and belong to the invalid area corresponding to the camera. After obtaining the historical trajectory data of the camera, the electronic device can determine the effective shooting area of ​​the camera in the video stream corresponding to the camera based on the distribution of trajectory points represented by the historical trajectory data. It should be noted that the effective shooting area corresponding to the camera can be one (in this case, the effective shooting area is the entire effective area of ​​the camera) or multiple (in this case, multiple effective shooting areas constitute the entire effective area of ​​the camera). Exemplarily, for the entire effective area of ​​camera A, the electronic device can divide the entire effective area into 3*3 effective shooting areas according to the preset segmentation method, and the associated cameras corresponding to each effective shooting area can be determined separately later.

[0111] In a possible implementation, step S303 may be implemented by the following steps (1) to (3):

[0112] (1) The overall shooting area of ​​the camera is divided into multiple sub-shooting areas.

[0113] (2) Among the multiple sub-shooting areas, the sub-shooting area that does not include the target trajectory point is deleted to obtain the target sub-shooting area.

[0114] (3) Merge the target sub-shooting areas to obtain the effective shooting area corresponding to the camera.

[0115] In the embodiment of the present application, the overall shooting area may refer to the entire visual range of the camera. The sub-shooting area may refer to the shooting area obtained by dividing the overall shooting area of ​​the camera. The target sub-shooting area may refer to the sub-shooting area of ​​the camera including the target trajectory point.

[0116] In this step, when the electronic device performs single-lens thermal analysis on the camera, with respect to the overall shooting area of ​​the camera, the electronic device can first divide the overall shooting area into multiple sub-shooting areas. For example, the overall shooting area of ​​the camera can be divided into countless 8*8 pixel sub-shooting areas. If the resolution of the camera is 1080*720, each row of the camera will be divided into 1080 / 8=135 sub-shooting areas, and each column will be divided into 720 / 8=90 sub-shooting areas. Finally, the entire lens is divided into 135*90=12150 sub-shooting areas.

[0117] Afterwards, the electronic device can determine the area that does not include the track point as an invalid area for multiple sub-shooting areas, and delete the invalid area from the sub-shooting area, so that the target sub-shooting area corresponding to the camera can be obtained. Specifically, based on the historical track data of a single camera, the electronic device can determine the number of track points in each sub-shooting area in order from the outside to the inside (from the surrounding to the middle), and can be determined based on the coordinate information of the track points and the image coordinates of the sub-shooting area. The sub-shooting area with a track number of 0 is deleted. If a sub-shooting area with a track number that is not 0 is encountered, the electronic device can stop the removal process of the current row or current column, and continue to delete the invalid area for the next row or next column. In this way, the electronic device can obtain the entire valid area corresponding to the camera.

[0118] For the entire effective area of ​​the camera, since the effective area as a whole may be an irregular polygon with a large number of edges, if the entire effective area is directly used as an effective shooting area for subsequent calculations, the amount of calculation is large, resulting in a more complicated calculation process for real-time tracking. On this basis, the electronic device can merge the target sub-shooting areas, specifically, the target sub-shooting areas can be evenly distributed to M*N (M*N is a positive integer) effective shooting areas to achieve segmentation of the entire effective area of ​​the camera. Exemplarily, the electronic device can specifically divide the number of target sub-shooting areas in each row by M, and the number of target sub-shooting areas in each column by N, so that all the number of target sub-shooting areas can be divided into each effective shooting area. In this way, the electronic device obtains multiple target sub-shooting areas by merging the target sub-shooting areas, and achieves segmentation of the entire effective area of ​​the camera, which can reduce the amount of calculation for subsequent real-time tracking and improve the real-time performance of target tracking.

[0119] For example, Figure 4 A schematic diagram of an effective shooting area of ​​a camera provided by an exemplary embodiment of the present application. Figure 4As shown, in the entire shooting area of ​​the camera, the video image of the camera may include invalid areas such as walls, which are inaccessible to the target foot points (black dots in the figure). The electronic device can first divide the entire shooting area to obtain multiple sub-shooting areas, and then identify and exclude from the image to the middle based on the historical trajectory data corresponding to a single camera and the various trajectory points in the single camera, delete the sub-shooting area that does not include the target trajectory point, and retain the target sub-shooting area including the trajectory point. The multiple target sub-shooting areas together constitute the effective area of ​​the camera. Afterwards, in order to facilitate the rapid calculation in the real-time tracking process, the electronic device can merge the target sub-shooting areas, divide the effective area of ​​the camera into M*N effective shooting areas (2*3 in the figure), and then analyze the corresponding cross-lens association relationship for each effective shooting area. In this way, the electronic device can reduce the search range of the target tracking process by identifying the effective area under a single camera, dividing the effective area, and subsequently establishing the target topological relationship between each effective shooting area and the associated camera, thereby improving the accuracy and real-time performance of target tracking.

[0120] S304: Determine the association relationship between the effective shooting area of ​​the camera and other cameras according to the effective shooting area and the trajectory session data, and obtain the target topological relationship corresponding to the multiple cameras.

[0121] In the embodiment of the present application, after determining the effective shooting area of ​​a single camera, the electronic device can determine the association relationship between the effective shooting area of ​​the camera and other cameras based on the trajectory session data. Specifically, the electronic device can determine the effective shooting area where the trajectory vanishes (i.e., the last trajectory point of the trajectory session data in the current camera) of the target trajectory session data is located, and based on the trajectory session data, determine the next camera that the same target reaches from the effective shooting, so that based on data such as the number of trajectories reaching other cameras, the electronic device can determine the association relationship between the effective shooting area and other cameras, and then obtain the target topological relationship corresponding to multiple cameras.

[0122] In a possible implementation, step S304 may be implemented by the following steps (4) to (6):

[0123] (4) In the effective shooting area of ​​the current camera, determine the target effective shooting area corresponding to the last track point of each track session data.

[0124] In an embodiment of the present application, the target effective shooting area may refer to the effective shooting area to which the last track point of the track session data in the camera belongs. Specifically, when the camera is performing track capture, if the target under the same lens does not leave the visible area of ​​the lens and does not split, then the track of this target is continuous and has time sequence. The camera can specifically report a track point to the electronic device at a fixed time period (for example, every 200 milliseconds), and the multiple track points reported for the same person include the same track identification (track ID), and each track point can specifically include the coordinate information and time information of the track point. On this basis, the electronic device can determine the target effective shooting area corresponding to the last track point of the track session data of the same target based on the coordinate information and time information of the track point in the current camera based on the track session data.

[0125] (5) According to the trajectory session data, determine the associated cameras corresponding to the effective shooting area of ​​the target, and determine the trajectory cross-lens time and trajectory quantity corresponding to each associated camera.

[0126] In the embodiment of the present application, the associated camera may refer to other cameras from which the target corresponding to the trajectory session data reaches the target effective shooting area. The track cross-lens time may refer to the time consumed by the trajectory session data from the target effective shooting area to the associated camera, that is, the time interval from the last track point of the previous camera to the first track point of the associated camera in the trajectory session data, and specifically may refer to the average track cross-lens time corresponding to the associated camera, etc. The number of tracks may refer to the number of target cross-lens tracks corresponding to the associated camera.

[0127] Specifically, after determining the target effective shooting area corresponding to the trajectory session data, the electronic device can further determine the associated camera corresponding to the target effective shooting area based on the trajectory session data. After that, the electronic device can analyze and count all the trajectory session data corresponding to the historical trajectory data, determine all the associated cameras corresponding to the target effective area, and determine the trajectory cross-lens time and trajectory number corresponding to each associated camera.

[0128] (6) Based on the trajectory cross-lens time and the number of trajectories, the target association probability between the target effective shooting area and the associated camera is determined.

[0129] (7) According to the target association probability, determine the target topological relationship corresponding to multiple cameras.

[0130] In the embodiment of the present application, the target association probability may refer to the cross-mirror probability between the target effective shooting area and the associated camera. The target association probability may be calculated based on the number of tracks and the track cross-mirror time, wherein the number of tracks may be used to calculate the ratio of the number of tracks of the associated camera to the total number of tracks of all associated cameras; the track cross-mirror time may correspond to different weights (or attenuation coefficients), and generally the longer the track cross-mirror time, the smaller the corresponding weight, that is, the longer the time it takes for the target to reach the associated camera from the target effective shooting area, the lower the possibility of the target reaching the associated camera from the target effective shooting area. The correlation relationship between the track cross-mirror time and the weight may be pre-set, and the embodiment of the present application does not limit this. After determining the target association probability of the associated camera corresponding to each target effective shooting area of ​​each camera, the electronic device may obtain the target topological relationship of multiple cameras based on the target association probability, and realize the determination of the connectivity relationship of multiple cameras, so that the user does not need to manually mark, saves human resources, and improves the real-time performance of target tracking.

[0131] Of course, the electronic device can also directly determine the target association probability based on the number of trajectories, without considering the weight caused by the cross-lens time of the trajectory, and the embodiments of the present application are not limited to this. Exemplarily, the associated cameras of the target effective shooting area A1 in camera A include associated cameras B and associated cameras C, wherein the number of trajectories of associated camera B is 250; the number of trajectories of associated camera C is 750. In this way, the target association probability corresponding to the associated camera B can be 250 / (250+750)*100%=25%, and the target association probability of the associated camera C can be 750 / (250+750)*100%=75%. On this basis, if the different weights caused by the trajectory cross-camera time are further considered, the target association probability can be further adjusted based on the weight. For example, the trajectory cross-camera time corresponding to the associated camera B is 90 seconds, and the corresponding weight adjustment value is X. The trajectory cross-camera time of the associated camera C is 30 seconds, and the corresponding weight adjustment value is Y. At this time, the target association probability of the associated camera B can be adjusted to 25%*X (or 25% minus X, etc.), and the target association probability of the associated camera C can be adjusted to 75%*Y (or 25% plus Y, etc.), where the values ​​of the weight adjustment values ​​X and Y can be flexibly set based on actual needs, and the embodiments of the present application are not limited to this.

[0132] It should be noted that in another possible implementation, the electronic device can also directly use the entire shooting area of ​​the camera as the effective shooting area for association analysis. In this case, there is no need to perform single-lens analysis and cutting of the camera. The associated cameras corresponding to the camera and the target association probabilities corresponding to each associated camera can be calculated later, and the target topological relationship between multiple cameras can be obtained based on the target association probability. In this way, the electronic device can improve the efficiency of determining the target topological relationship and simplify the calculation process by directly analyzing and associating the entire shooting area of ​​the camera.

[0133] For example, Figure 5 A schematic diagram of trajectory statistics of cross-mirror relationship of trajectory session data provided by an exemplary embodiment of the present application. Figure 5 As shown, the starting camera may refer to the single-lens camera currently being analyzed, which includes 2*3 effective shooting areas, specifically including effective shooting areas 1-1, 1-2, 1-3, 2-1, 2-2, and 2-3. The target camera may refer to the associated camera corresponding to the effective shooting area. The total duration may refer to the total cross-lens time of all trajectories from the effective shooting area to the associated camera. The number of trajectories may refer to the number of trajectories from the effective shooting area to the target camera. The average duration may refer to the cross-lens time of the trajectory from the effective shooting area to the associated camera.

[0134] Specifically, the electronic device determines the target effective shooting area corresponding to the last track point of each track session data in a single camera during single-lens analysis based on the track session data of the same target, and then determines the associated camera corresponding to the target effective shooting area based on the track session data, so that all track session data in the preset historical period are analyzed and counted to obtain the number of tracks and track cross-lens time of each effective shooting area reaching the associated camera, and then the target association probability of the effective shooting area to the associated camera can be calculated based on the number of tracks and the track cross-lens time, and then the target topological relationship between multiple cameras can be obtained, so that no manual intervention is required, the target topological relationship is automatically determined, human resources are saved, and the topological relationship between cameras can be updated based on the periodic changes of historical track data, ensuring the validity of the topological relationship. It should be noted that for associated cameras with a track number less than a preset threshold, such as a track number less than 5, the electronic device can determine that the probability of association between the effective shooting area and the associated camera is small, and the associated camera can be discarded, which can save computing resources and avoid interference from invalid data.

[0135] S305, collecting target feature information of each target in the first camera; when the target feature information matches the target feature information corresponding to the target person in the preset database, determining the target to which the target feature information belongs as the target to be tracked.

[0136] In an embodiment of the present application, the preset database may refer to a pre-set database of specific personnel, the target personnel may refer to a specific personnel, and the target feature information may refer to the feature information corresponding to the specific personnel. Specifically, in offline retail scenarios, for specific personnel with risky behaviors, supermarket staff can store the target feature information of the specific personnel in the preset database, which can be used to identify and judge the target personnel later. The electronic device collects the target feature information of each target in the real-time image stream of the first camera, and performs feature matching on the target feature information of each target with the target feature information of the target personnel in the preset database. If the preset database stores target feature information that matches the target feature information, the electronic device can determine that the target corresponding to the target feature information is a specific person, and the specific person can be used as a target to be tracked later. In this way, the electronic device can accurately determine the target to be tracked by comparing and matching the target feature information with the target feature information, which can improve the accuracy and rationality of target tracking.

[0137] S306: Determine feature information and location information corresponding to the target to be tracked.

[0138] S307, determining a target effective shooting area corresponding to the target to be tracked in the first camera according to the position information; and determining a target associated camera corresponding to the target to be tracked based on the target topological relationship and the target effective shooting area.

[0139] In an embodiment of the present application, the target effective shooting area may refer to the effective shooting area to which the last trajectory point of the target to be tracked in the first camera belongs. When the electronic device performs real-time tracking of the target to be tracked, the characteristic information and position information of the target to be tracked may be determined in the first camera. Afterwards, the electronic device may determine the target effective shooting area corresponding to the target to be tracked in the first camera based on the position information, and then search in the target topological relationship based on the target effective shooting area to determine the target-associated camera after the first camera for the target to be tracked, and then search and match in the target-associated camera to achieve real-time tracking of the target to be tracked.

[0140] S308: Determine the target association probability corresponding to each target-associated camera according to the target topological relationship.

[0141] S309 , searching for feature information in each target-associated camera in descending order of target-associated probability.

[0142] S310: When target feature information in the target-associated camera matches feature information of the target to be tracked, determine real-time position information of the target to be tracked.

[0143] In an embodiment of the present application, after determining the target-associated camera, the electronic device can further determine the target association probability corresponding to each target-associated camera, and search for the feature information of the target to be tracked in each target-associated camera in order of the target association probability from high to low. When the similarity between the target feature information in the target-associated camera and the feature information of the target to be tracked is greater than a preset similarity threshold, the electronic device can determine that the target to be tracked has arrived at the target-associated camera, and can determine the real-time position information of the target to be tracked based on the trajectory point information reported by the target-associated camera, thereby realizing real-time cross-mirror tracking of the target to be tracked.

[0144] For example, Figure 6 A schematic diagram of real-time tracking provided by an exemplary embodiment of the present application. Figure 6 As shown, the electronic device determines the target to be tracked based on the target feature information in the first camera, and determines the target to be tracked ( Figure 6 The electronic device can save the feature information and location information of the target to be tracked in a track session data. As the target moves, the feature information and location information will also change in real time, and all of this information will be saved in the session. When the target to be tracked disappears in the first camera, the electronic device can determine the target effective shooting area to which the last track point of the target to be tracked belongs based on the location information of the target to be tracked, and search in the target topological relationship to determine the target-associated camera corresponding to the target effective shooting area and the target association probability corresponding to each target-associated camera. After that, the electronic device can search and match the feature information in each target-associated camera according to the level of the target association probability, and specifically determine that the target feature information of each target in the target-associated camera is matched with the feature information of the target to be tracked for similarity. When the similarity is greater than the preset similarity threshold, the electronic device can determine that the target to be tracked is found, and then the real-time location information of the target to be tracked can be determined. In this way, the electronic device can achieve accurate and real-time tracking of a specific person based on the location information, feature information of the target to be tracked, and the target topological relationship between multiple cameras determined in advance.

[0145] It should be understood that in the various embodiments of the present application, the size of the serial number of each process does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0146] Based on any one of the above embodiments, Figure 7 A schematic diagram of a technology link for real-time target tracking provided by an exemplary embodiment of the present application. Figure 7As shown, the electronic device obtains the historical trajectory data of each camera (camera 1 to camera n), performs thermal analysis on each camera based on the historical trajectory data, identifies the effective area in a single camera, and divides the effective area to obtain the effective shooting area of ​​a single camera. Afterwards, the electronic device can merge the historical trajectory data based on REID to obtain the trajectory session data of each target, and determine the effective shooting area of ​​the target where the last trajectory point of the trajectory session data is located under a single lens, and further determine the next associated camera that the target reaches from the target effective shooting area based on the trajectory session data, so as to obtain the associated cameras of each effective shooting area, and also obtain the number of trajectories, trajectory cross-mirror time and other data corresponding to each associated camera. Then, the electronic device can determine the target association probability of the effective shooting area with each associated camera based on the number of trajectories, trajectory cross-mirror time and other data, and then determine the target topological relationship between multiple cameras. After determining the target topological relationship between multiple cameras, the electronic device can realize real-time tracking of the target to be tracked based on the target topological relationship, combined with the feature information and location information of the target to be tracked.

[0147] In related technologies, when supermarket staff in offline retail scenarios track and monitor specific people, they usually mark the direction of each camera and the surrounding cameras that can be reached in the system, and then track them based on the manually calibrated camera position relationship. This tracking method has the following disadvantages:

[0148] First, the accuracy of real-time tracking is not high. Since there are usually a large number of people in offline scenes, there are too many similar people at the same time. For example, they wear the same black clothes or pants, which can easily cause misidentification. It is also easy to have tracking errors during real-time tracking. Second, the installation positions of the lenses in the offline scenes are not uniform, and the features captured for the target will be various, resulting in greater interference in feature extraction and a higher probability of misidentification. Third, the speed and direction of movement of each person are difficult to predict, and the range of a camera is huge. Depending on the installation position of the camera, this range varies greatly. The actual physical distance between the upper left corner and the lower right corner of the same camera is even more than 50 meters. The related technology uses a simple manual calibration method for cameras, which has large errors and can easily route to the wrong next camera, resulting in failure to track a specific person.

[0149] In the embodiment of the present application, the electronic device can calculate the trajectory session data of all the targets in the field based on the historical trajectory data of a large number of targets in the field, and can determine the target topological relationship of multiple cameras based on the single-lens thermal analysis and cross-lens relationship analysis. In addition, since the effective shooting area of ​​a single camera is divided during the single-lens thermal analysis process, based on the effective shooting area and the target topological relationship, the electronic device can quickly predict the next target-associated camera that the current target may reach in a certain effective shooting area, which can greatly reduce the search range, thereby improving the accuracy of real-time tracking and improving the user experience of the staff.

[0150] In addition, in the embodiment of the present application, the electronic device collects the trajectory information of the whole field, calculates the thermal force of each area of ​​the camera, and combines the REID trajectory results to determine the topology and probability relationship of the regional grid of the lens to the next camera, and then the target topology relationship between multiple cameras can be obtained, without the need for manual marking by the user, providing an automated solution for real-time tracking, greatly reducing the workload of manual participation, with a wider range of uses, and can also achieve large-scale implementation. Subsequently, the edge end (electronic device) can realize real-time tracking of the target to be tracked based on the feature information and location information of the target to be tracked, and based on the target topology relationship, it also improves the accuracy of real-time tracking.

[0151] Figure 8 A schematic diagram of a real-time target tracking device provided by an exemplary embodiment of the present application is shown in FIG. Figure 8 , the target real-time tracking device 80 comprises:

[0152] An acquisition module 81 is used to acquire historical trajectory data corresponding to multiple cameras, and determine the target topological relationship corresponding to the multiple cameras according to the historical trajectory data;

[0153] A first determination module 82, configured to determine a target to be tracked based on the target feature information in the first camera, and determine feature information and position information corresponding to the target to be tracked;

[0154] A second determination module 83 is used to determine a target-associated camera corresponding to the first camera based on the position information and the target topological relationship;

[0155] The third determination module 84 is used to determine the real-time position information of the target to be tracked based on the feature information in the target associated camera.

[0156] In a possible implementation, the acquisition module 81 is specifically configured to:

[0157] Obtain the track point information reported by each camera; the track point information includes the coordinate information and time information of each track point;

[0158] Get the track point information within the preset historical period and obtain the historical track data.

[0159] In a possible implementation, the acquisition module 81 is specifically configured to:

[0160] The historical trajectory data corresponding to multiple cameras are merged to obtain the trajectory session data corresponding to each target;

[0161] For each camera, determine the effective shooting area corresponding to the camera based on the camera's historical trajectory data;

[0162] According to the effective shooting area and the trajectory session data, the association relationship between the effective shooting area of ​​the camera and other cameras is determined, and the target topological relationship corresponding to the multiple cameras is obtained.

[0163] In a possible implementation, the acquisition module 81 is specifically configured to:

[0164] The overall shooting area of ​​the camera is divided into multiple sub-shooting areas;

[0165] In the plurality of sub-shooting areas, a sub-shooting area not including a target track point is deleted to obtain a target sub-shooting area;

[0166] The target sub-shooting areas are merged to obtain the effective shooting area corresponding to the camera.

[0167] In a possible implementation, the acquisition module 81 is specifically configured to:

[0168] In the effective shooting area of ​​the current camera, determine the target effective shooting area corresponding to the last track point of each track session data;

[0169] According to the trajectory session data, determine the associated cameras corresponding to the effective shooting area of ​​the target, and determine the trajectory cross-lens time and trajectory quantity corresponding to each associated camera;

[0170] Based on the trajectory cross-lens time and the number of trajectories, the target association probability between the target effective shooting area and the associated camera is determined;

[0171] According to the target association probability, the target topological relationship corresponding to multiple cameras is determined.

[0172] In a possible implementation manner, the first determining module 82 is specifically configured to:

[0173] Collecting target feature information of each target in the first camera;

[0174] When the target characteristic information matches the target characteristic information corresponding to the target person in the preset database, the target to which the target characteristic information belongs is determined as the target to be tracked.

[0175] In a possible implementation manner, the second determining module 83 is specifically configured to:

[0176] Determine, according to the position information, a target effective shooting area corresponding to the target to be tracked in the first camera;

[0177] Based on the target topological relationship and the target effective shooting area, the target associated camera corresponding to the target to be tracked is determined.

[0178] In a possible implementation manner, the third determining module 84 is specifically configured to:

[0179] According to the target topological relationship, the target association probability corresponding to each target-associated camera is determined;

[0180] Search for feature information in each target-associated camera in descending order of target-associated probability;

[0181] When target feature information in the target-associated camera matches feature information of the target to be tracked, the real-time position information of the target to be tracked is determined.

[0182] The target real-time tracking device 80 provided in the embodiment of the present application can execute the technical solution shown in the above method embodiment, and its implementation principle and beneficial effects are similar, which will not be repeated here.

[0183] Fig. 9 For a schematic diagram of the structure of an electronic device provided by an exemplary embodiment of the present application, see Fig. 9 The electronic device 90 may include a processor 91 and a memory 92. Exemplarily, the processor 91 and the memory 92 are interconnected via a bus 93.

[0184] Memory 92 stores computer executable instructions;

[0185] The processor 91 executes the computer-executable instructions stored in the memory 92, so that the processor 91 executes the target real-time tracking method shown in the above method embodiment.

[0186] Accordingly, an embodiment of the present application provides a computer-readable storage medium, in which computer-executable instructions are stored. When the computer-executable instructions are executed by a processor, they are used to implement the target real-time tracking method of the above method embodiment.

[0187] Accordingly, an embodiment of the present application may also provide a computer program product, including a computer program, which, when executed by a processor, can implement the target real-time tracking method shown in the above method embodiment.

[0188] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0189] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0190] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0191] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0192] In a typical configuration, a computing device includes one or more processors, input / output interfaces, network interfaces, and memory.

[0193] The memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0194] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0195] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0196] The above is only an embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

Claims

1. A real-time target tracking method, characterized in that: include: Get historical trajectory data corresponding to multiple cameras; Merging the historical trajectory data corresponding to the multiple cameras to obtain trajectory session data corresponding to each target; For each camera, based on the historical trajectory data of the camera, determining the effective shooting area corresponding to the camera; Determine, according to the effective shooting area and the trajectory session data, the association relationship between the effective shooting area of ​​the camera and other cameras, and obtain the target topological relationship corresponding to the multiple cameras; Determining a target to be tracked based on the target feature information in the first camera, and determining feature information and position information corresponding to the target to be tracked; Based on the location information and the target topological relationship, determining a target-associated camera corresponding to the first camera; In the target-associated camera, the real-time position information of the target to be tracked is determined based on the feature information.

2. The method according to claim 1, characterized in that: The obtaining of historical trajectory data corresponding to the plurality of cameras includes: Acquire the track point information reported by each camera; the track point information includes the coordinate information and time information of each track point; The track point information within a preset historical period is obtained to obtain the historical track data.

3. The method according to claim 1, characterized in that: The determining, based on the historical trajectory data of the camera, a valid shooting area corresponding to the camera includes: Dividing the overall shooting area of ​​the camera into multiple sub-shooting areas; Among the multiple sub-shooting areas, deleting the sub-shooting area that does not include the target trajectory point, to obtain the target sub-shooting area; The target sub-shooting areas are merged to obtain an effective shooting area corresponding to the camera.

4. The method according to claim 1, characterized in that The determining, according to the effective shooting area and the trajectory session data, the association relationship between the effective shooting area of ​​the camera and other cameras, and obtaining the target topological relationship corresponding to the multiple cameras, includes: In the effective shooting area of ​​the current camera, determine the target effective shooting area corresponding to the last track point of each track session data; Determine, according to the trajectory conversation data, the associated cameras corresponding to the effective shooting area of ​​the target, and determine the trajectory cross-lens time and the number of trajectories corresponding to each of the associated cameras; Determine the target association probability between the target effective shooting area and the associated camera based on the trajectory cross-lens time and the number of trajectories; According to the target association probability, the target topological relationship corresponding to the multiple cameras is determined.

5. The method according to claim 1, characterized in that The step of determining the target to be tracked based on the target feature information in the first camera includes: Collecting target feature information of each target in the first camera; In the case where the target characteristic information matches the target characteristic information corresponding to the target person in the preset database, the target to which the target characteristic information belongs is determined as the target to be tracked.

6. The method according to claim 4, characterized in that The determining, based on the location information and the target topological relationship, a target-associated camera corresponding to the first camera includes: Determine, according to the position information, a target effective shooting area corresponding to the target to be tracked in the first camera; Based on the target topological relationship and the target effective shooting area, a target-associated camera corresponding to the target to be tracked is determined.

7. The method according to claim 1 or 6, characterized in that: The step of determining the real-time position information of the target to be tracked based on the feature information in the target-associated camera includes: Determine the target association probability corresponding to each target-associated camera according to the target topological relationship; Searching for the feature information in each target-associated camera in descending order of target-associated probability; In the case that target feature information exists in the target-associated camera and matches feature information of the target to be tracked, the real-time position information of the target to be tracked is determined.

8. A target real-time tracking device, characterized in that: include: An acquisition module, used to acquire historical trajectory data corresponding to multiple cameras, merge the historical trajectory data corresponding to the multiple cameras, and obtain trajectory session data corresponding to each target; For each camera, based on the historical trajectory data of the camera, determining the effective shooting area corresponding to the camera; Determine, according to the effective shooting area and the trajectory session data, the association relationship between the effective shooting area of ​​the camera and other cameras, and obtain the target topological relationship corresponding to the multiple cameras; A first determination module, configured to determine a target to be tracked based on target feature information in a first camera, and determine feature information and position information corresponding to the target to be tracked; A second determination module, configured to determine a target-associated camera corresponding to the first camera based on the location information and the target topological relationship; The third determination module is used to determine the real-time position information of the target to be tracked based on the feature information in the target-associated camera.

9. An electronic device, characterized in that: include: Memory and processor; The memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory, so that the processor executes the target real-time tracking method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the target real-time tracking method according to any one of claims 1 to 7.

11. A computer program product, characterized in that The invention comprises a computer program, which, when executed by a computer, implements the target real-time tracking method according to any one of claims 1 to 7.

Images