Target trajectory positioning method and device, storage medium and computer equipment

Abstract

According to the target track positioning method and device, the storage medium and the computer equipment, after it is detected that the person enters the target area, the topological relation graph of the cameras in the target area is obtained, the cameras are only deployed at all the path forks in the target area, and therefore the hardware deployment cost can be reduced, and the user experience is improved. And then determining the identity of the person through a person image collected by a camera with the monitoring range covering the person, and determining and updating a real-time trajectory positioning map of the person based on the identity of the person, thereby improving the accuracy of continuous updating of the trajectory. After a person leaves the monitoring range of the camera, the camera communicated with the camera path in the topological relation graph is monitored in real time, blind searching of all cameras is avoided, the positioning efficiency is improved, and when it is monitored that a person appears in the monitoring range of the camera, the corresponding camera is switched to continue to perform positioning updating on the person track, so that the positioning accuracy of the person is improved. And continuous and seamless tracking of the person is realized until the person leaves the target area.

Description

Technical Field

[0001] The present application relates to the technical field of trajectory positioning, and in particular, to a method, apparatus, storage medium, and computer device for target trajectory positioning. Background Art

[0002] With the increasing demands of modern security, logistics, and personnel management, target trajectory positioning technology has become a key means to ensure security and improve efficiency. In the early days, positioning technology mostly relied on outdoor positioning systems such as GPS. However, in indoor, underground, or dense areas, GPS signals are limited, making it difficult to achieve accurate positioning. Against this background, various target trajectory positioning methods, such as RFID-based positioning systems, WiFi-based indoor positioning, and full-coverage video surveillance systems, have gradually developed.

[0003] However, although the existing technologies meet the indoor positioning requirements to a certain extent, both RFID readers and dense camera networks require a large amount of capital investment, and the equipment maintenance costs are not low. In addition, they have many usage restrictions. For example, RFID requires personnel to carry tags, and WiFi positioning depends on the cooperation of user devices, making it difficult to achieve true unobtrusive tracking. In short, the existing target trajectory positioning methods are still difficult to meet the requirements of complex and changing scenarios, and it is often difficult to balance positioning accuracy and real-time performance, resulting in low personnel management efficiency. Summary of the Invention

[0004] The purpose of the present application is to at least solve one of the above technical defects, especially the technical defect that the existing target trajectory positioning methods are still difficult to meet the requirements of complex and changing scenarios, and it is often difficult to balance positioning accuracy and real-time performance, resulting in low personnel management efficiency.

[0005] The present application provides a method for target trajectory positioning, and the method includes:

[0006] After detecting that a person enters a target area, obtain the topological relationship diagram of the cameras in the target area, and mark the cameras whose monitoring ranges cover the person as target cameras; wherein, cameras are deployed at each path bifurcation in the target area;

[0007] Collect images of the person through the target cameras to obtain person images, and determine the identity of the person in the person images based on the personnel information database;

[0008] Based on the determined identity of the person, determine and update the real-time trajectory positioning map of the person, and after the person leaves the monitoring range of the target cameras, determine at least one camera that is path-connected to the target cameras according to the topological relationship diagram, and mark it as a candidate camera;

[0009] Monitor the candidate cameras in real time. When it is detected that the person appears within the monitoring range of a candidate camera, mark the corresponding candidate camera as a new target camera, and return to collect an image of the person through the target camera to obtain a person image and its subsequent steps until it is detected that the person leaves the target area.

[0010] Optionally, the obtaining of the topological relationship graph of the cameras in the target area includes:

[0011] Obtain the path structure of the target area and the installation positions of each camera;

[0012] Analyze the physical connectivity between each installation position according to the path structure, and establish an accessibility matrix between each camera according to the analysis result;

[0013] Based on the accessibility matrix, determine the connectivity paths between each adjacent and connected camera, and construct a topological relationship graph according to each camera and its corresponding connectivity path.

[0014] Optionally, the determining of the identity of the person in the person image based on the person information database includes:

[0015] Obtain the person image from the target camera through a video recorder, and extract the person features of the person image;

[0016] Perform feature matching on the person features in the person information database, and determine the identity of the person in the person image according to the matching result;

[0017] Among them, multiple person information is registered in advance in the person information database, and each person information includes a person identity and person features.

[0018] Optionally, the determining of the identity of the person in the person image according to the matching result includes:

[0019] Judge whether the matching result is empty;

[0020] If so, generate a temporary ID, bind the temporary ID to the person features and store it in the person information database, and use the temporary ID as the identity of the person in the person image;

[0021] If not, directly obtain the identity ID in the matching result as the identity of the person in the person image.

[0022] Optionally, the determining and updating of the real-time trajectory positioning map of the person based on the person identity includes:

[0023] Determine whether there is a real-time trajectory positioning map of the person's identity in the target area;

[0024] If so, obtain the real-time trajectory positioning map, and update the moving trajectory in the real-time trajectory positioning map according to the installation position of the target camera;

[0025] If not, create a real-time trajectory positioning map of the person's identity in the target area, and add the moving trajectory of the person to the real-time trajectory positioning map according to the installation position of the target camera.

[0026] Optionally, the determining at least one camera connected to the target camera path according to the topological relationship map and marking it as a candidate camera includes:

[0027] Determine the connected path corresponding to the target camera in the topological relationship map, and mark the camera at the other end of each connected path as a candidate camera.

[0028] Optionally, the method further includes:

[0029] After real-time monitoring of the candidate cameras, record the monitoring duration of the candidate cameras until it is detected that the person appears in the monitoring range of a candidate camera;

[0030] If the monitoring duration exceeds a preset threshold, stop the real-time monitoring of the candidate cameras, perform trajectory analysis on the real-time trajectory positioning map, and re-determine candidate cameras according to the analysis results and perform real-time monitoring.

[0031] This application also provides a target trajectory positioning device, including:

[0032] A personnel detection module, configured to, after detecting that a person enters the target area, obtain the topological relationship map of the cameras in the target area, and mark the camera whose monitoring range covers the person as the target camera; wherein, cameras are deployed at each path bifurcation in the target area;

[0033] An identity determination module, configured to collect an image of the person through the target camera to obtain a person image, and determine the identity of the person in the person image based on the personnel information database;

[0034] A trajectory update module, configured to determine and update the real-time trajectory positioning map of the person based on the personnel identity, and after the person leaves the monitoring range of the target camera, determine at least one camera connected to the target camera path according to the topological relationship map and mark it as a candidate camera;

[0035] The cross-camera monitoring module is used to monitor the candidate cameras in real time. When it detects that the person appears within the monitoring range of a candidate camera, it marks the corresponding candidate camera as a new target camera, and returns to perform image acquisition of the person through the target camera to obtain a person image and subsequent steps until it detects that the person leaves the target area.

[0036] The present application also provides a storage medium, in which computer-readable instructions are stored. When the computer-readable instructions are executed by one or more processors, the one or more processors are caused to execute the steps of the target trajectory positioning method as described in any one of the above embodiments.

[0037] The present application also provides a computer device, including: one or more processors, and a memory;

[0038] The memory stores computer-readable instructions. When the computer-readable instructions are executed by the one or more processors, the steps of the target trajectory positioning method as described in any one of the above embodiments are executed.

[0039] As can be seen from the above technical solutions, the embodiments of the present application have the following advantages:

[0040] For the target trajectory positioning method, device, storage medium and computer device provided by the present application, after detecting that a person enters a target area, it can obtain a topological relationship map of cameras in the target area, and mark the cameras whose monitoring ranges cover the person as target cameras; among them, cameras are only deployed at each path bifurcation in the target area, so as to ensure comprehensive tracking of the person's trajectory while reducing the hardware deployment cost. Then, it can perform image acquisition of the person through the target camera to obtain a person image, and determine the identity of the person in the person image based on the personnel information database. Then, it can determine and update the real-time trajectory positioning map of the person based on the person's identity, thereby improving the accuracy of continuous trajectory update. After the person leaves the monitoring range of the target camera, it can determine at least one camera that is path-connected to the target camera according to the topological relationship map, and mark it as a candidate camera, avoiding blindly searching all cameras and improving the positioning efficiency; finally, it can monitor the candidate cameras in real time, and when it detects that the person appears within the monitoring range of a candidate camera, it marks the corresponding candidate camera as a new target camera to continue to perform positioning and update of the person's trajectory through the target camera, realizing continuous and seamless tracking of the person until the person leaves the target area, ensuring the integrity of trajectory positioning. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] To more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the accompanying drawings required for the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0042] Figure 1 Schematic flowchart of a target trajectory positioning method provided by an embodiment of the present application;

[0043] Figure 2 Schematic flowchart of a topological relationship graph construction process provided by an embodiment of the present application;

[0044] Figure 3 Schematic structural diagram of a target trajectory positioning device provided by an embodiment of the present application;

[0045] Figure 4 Schematic internal structure diagram of a computer device provided by an embodiment of the present application. Detailed implementation manners

[0046] The following will clearly and completely describe the technical solutions in the embodiments of the present application with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, rather than all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present application.

[0047] The prior art meets the indoor positioning requirements to a certain extent. However, whether it is an RFID reader or a dense camera network, a large amount of capital investment is required, and the equipment maintenance cost is not low; in addition, there are many usage restrictions. For example, RFID requires personnel to carry tags, and WiFi positioning depends on the cooperation of user devices, making it difficult to achieve true unobtrusive tracking. All in all, the existing target trajectory positioning methods are still difficult to meet the complex and changeable scenario requirements, and it is often difficult to balance the positioning accuracy and real-time performance, resulting in low personnel management efficiency.

[0048] Based on this, the present application proposes the following technical solutions. For details, please refer to the following:

[0049] In one embodiment, as Figure 1 shown, Figure 1 Schematic flowchart of a target trajectory positioning method provided by an embodiment of the present application; The present application provides a target trajectory positioning method, which specifically includes the following:

[0050] S110: After detecting that a person has entered the target area, obtain the topological relationship diagram of the cameras in the target area, and mark the cameras whose monitoring ranges cover the person as target cameras; each path bifurcation in the target area is equipped with a camera.

[0051] In this step, after the computer device detects that a person has entered the target area, it can obtain the topological relationship diagram of the cameras in the target area, and mark the cameras whose monitoring ranges cover the person as target cameras; each path bifurcation in the target area is only equipped with a camera, so as to ensure comprehensive tracking of the person's trajectory while reducing the hardware deployment cost.

[0052] Among them, since each path bifurcation in the target area is only equipped with a camera, its topological relationship diagram refers to a graphical structure formed by taking the cameras in the target area as nodes and the paths that people can pass through as edges, and is a graphical structure used to describe the spatial layout and connectivity relationship between the cameras in the target area.

[0053] It can be understood that in order to reduce the system hardware deployment cost while ensuring comprehensive tracking of the person's trajectory, the target area adopts a scheme of only arranging cameras at each key path bifurcation in the design, avoiding dense deployment on all paths. This kind of layout method can not only effectively reduce the number of cameras and the installation and maintenance cost, but also, with the help of the traffic logic relationship provided by the topological relationship diagram, still realize the continuous recognition and accurate prediction of the person's trajectory.

[0054] Specifically, when the computer device can confirm that a person has entered the target area through the received trigger signal, such as the record of the access control system, the induction of the infrared sensor, etc., at this time the person is located in the detection area at the entrance of the target area, and the monitoring range of the camera deployed in this detection area can cover the person. Therefore, the computer device can mark this camera as the target camera, and starting from this target camera, combine the position layout relationship of each camera in the topological relationship diagram to perform trajectory positioning and tracking of the person within the target area.

[0055] S120: Collect an image of the person through the target camera to obtain a person image, and determine the identity of the person in the person image based on the personnel information database.

[0056] In this step, after determining the target camera through step S110, the computer device can collect an image of the person through the target camera to obtain a person image, and determine the identity of the person in the person image based on the personnel information database. By binding the personnel identity, the situation of incorrect target tracking during the trajectory positioning process can be avoided.

[0057] Specifically, after locking the target camera, the computer device can directly capture real-time images of the person through the target camera, so as to obtain clear person images, thereby avoiding the blind capture of person images by cameras in the target area. Then, the computer device can perform front-end processing on the obtained person images and then perform person feature recognition, improving the accuracy of person feature recognition by improving the quality of the person images. After identifying the person features, the computer device can match the corresponding person from the pre-constructed person information database based on the person features and extract their person identity.

[0058] It can be understood that confirming the person's identity before determining the person's trajectory positioning can enable the computer device to maintain the uniqueness and continuity of the target object throughout the process of person trajectory tracking, avoiding tracking errors or target drift problems caused by situations such as multiple people being similar, camera switching, and image occlusion.

[0059] S130: Determine and update the real-time trajectory positioning map of the person based on the person's identity, and after the person leaves the monitoring range of the target camera, determine at least one camera that is path-connected to the target camera according to the topological relationship map and mark it as a candidate camera.

[0060] In this step, after determining the person's identity through step S120, the computer device can determine and update the real-time trajectory positioning map of the person based on the person's identity, thereby improving the accuracy of continuous trajectory update; when the person leaves the monitoring range of the target camera, the computer device can determine at least one camera that is path-connected to the target camera according to the topological relationship map and mark it as a candidate camera, avoiding blindly searching all cameras and improving the positioning efficiency.

[0061] Among them, the real-time trajectory positioning map of the present application refers to a visual trajectory map or data structure dynamically maintained by the computer device for each person entering the target area, reflecting the position change of the person in the target area. It records the position and time information of the corresponding person appearing under different cameras in a time series manner, forming a complete and continuous movement trajectory.

[0062] Specifically, after the computing device identifies the person's identity through the target camera, it can use the person's identity as an index to update the trajectory positioning of the corresponding person in the target area in real time, forming a real-time trajectory map. Through the close binding of the person's identity and the real-time trajectory positioning map, the continuous tracking accuracy and stability of the person's movement path can be significantly improved, especially in cross-camera recognition and fast-switching scenarios, effectively reducing trajectory breakpoints or incorrect records caused by problems such as target misjudgment and tracking loss.

[0063] Furthermore, when the computer device detects that the person has left the monitoring range of the target camera, it will not blindly conduct a global search among all cameras, but will immediately search for adjacent cameras that have a path connectivity relationship with the camera based on a pre-established topological relationship graph. These cameras represent the next location that the person may go to in the current physical space. Therefore, these cameras can be automatically marked as candidate cameras for targeted monitoring, thereby narrowing the search range for the person and reducing computing resource consumption.

[0064] S140: Monitor the candidate cameras in real time, and when a person is detected within the monitoring range of the candidate cameras, mark the corresponding candidate cameras as new target cameras, and return to capture images of the person through the target cameras to obtain person images and subsequent steps until it is detected that the person leaves the target area.

[0065] In this step, after the candidate cameras are determined in step S130, the computer device can monitor each candidate camera in real time, and when a person appears in the monitoring range of the candidate camera, the corresponding candidate camera is marked as a new target camera, so as to continue to locate and update the trajectory of the person through the target camera, thereby achieving continuous and seamless tracking of the person until the person leaves the target area, thereby ensuring the integrity of the trajectory positioning.

[0066] Specifically, the computer device can prioritize candidate cameras and continuously monitor whether the target person reappears within its monitoring range. Once a person is detected in a candidate camera, the computer device can immediately mark the candidate camera as a new target camera and perform image acquisition and person detection to complete the seamless connection of the tracked object, and continue to update the real-time trajectory positioning map of the person based on the newly marked target camera.

[0067] It is understandable that as people continue to appear within the monitoring range of different cameras, the computer equipment can continuously update the real-time trajectory positioning map of the target camera and the person, and achieve continuous and accurate tracking of the target. The entire process does not require human intervention, and the decision switching can be completed within milliseconds, ensuring the efficient response and tracking continuity of the real-time trajectory positioning map. Until the person finally leaves the preset target area, the computer equipment can mark his status as leaving and stop subsequent tracking records, thus forming a complete, closed-loop, continuous and uninterrupted trajectory link, ensuring the high integrity and data accuracy of the entire trajectory positioning process.

[0068] In the above embodiments, after detecting that a person enters the target area, a topological relationship graph of the cameras in the target area can be obtained, and the cameras whose monitoring ranges cover the person are marked as target cameras. Among them, cameras are only deployed at each path bifurcation in the target area, so as to ensure comprehensive tracking of the person's trajectory while reducing the hardware deployment cost. Then, images of the person can be collected by the target cameras to obtain person images, and the identity of the person in the person images can be determined based on the person information database. Then, the real-time trajectory positioning graph of the person can be determined and updated based on the person identity, thereby improving the accuracy of continuous trajectory update. After the person leaves the monitoring range of the target camera, at least one camera that is path-connected to the target camera can be determined according to the topological relationship graph and marked as a candidate camera, avoiding blindly searching all cameras and improving the positioning efficiency. Finally, the candidate cameras can be monitored in real time, and when it is detected that a person appears in the monitoring range of a candidate camera, the corresponding candidate camera is marked as a new target camera, so as to continue to locate and update the person's trajectory through the target camera, realizing continuous and seamless tracking of the person until the person leaves the target area, ensuring the integrity of trajectory positioning.

[0069] In one embodiment, as Figure 2 shown, Figure 2 is a schematic flowchart of a process for constructing a topological relationship graph provided by an embodiment of the present application; Figure 2 In it, the process of obtaining the topological relationship graph of the cameras in the target area in step S110 may include:

[0070] S111: Obtain the path structure of the target area and the installation positions of each camera.

[0071] S112: Analyze the physical connectivity between each installation position according to the path structure, and establish a reachability matrix between each camera according to the analysis result.

[0072] S113: Determine the connection paths between each adjacent and connected camera based on the reachability matrix, and construct a topological relationship graph according to each camera and its corresponding connection path.

[0073] In this embodiment, before obtaining the topological relationship graph, the computer device can first obtain the path structure of the target area and the installation positions of each camera, then analyze the physical connectivity between each installation position according to the path structure, and establish a reachability matrix between each camera according to the analysis result. Furthermore, the connection paths between each adjacent and connected camera can be determined based on the reachability matrix, and a topological relationship graph can be constructed according to each camera and its corresponding connection path.

[0074] Among them, the reachability matrix refers to a two-dimensional matrix structure used to describe whether there is a path connection relationship between cameras in the target area, and is used to express the indirect physical accessibility between any two cameras. Through the reachability matrix, the computer device can systematically analyze the connectivity in the entire camera network, and then generate a topological connection diagram between cameras.

[0075] Specifically, since the path information includes spatial information such as the connection method of the path, the intersection points between each path, the bifurcation points and their directions, and the installation position of the camera points to the physical coordinates or nodes corresponding to each device in the actual scene, the computer device can comprehensively analyze the physical connectivity of the entire target area based on this information, and then can determine whether there is a path for personnel to pass between different camera nodes. Finally, a reachability matrix is established based on the analysis results. Here, the reachability matrix can use the camera as the row and column coordinates, and each element represents whether there is a connected path between two cameras. If there is, it is assigned a value of 1, and if not, it is assigned a value of 0.

[0076] Therefore, through this reachability matrix, the computer device can determine each adjacent and connected camera, as well as the connected path between them, and then construct a topological relationship diagram of the cameras in the target area with the camera as the node and the connected path as the edge, providing a basic support for subsequent intelligent analysis functions such as target tracking, camera scheduling, and path prediction.

[0077] In one embodiment, the process of determining the personnel identity of the personnel in the person image based on the personnel information database in step S120 may include:

[0078] S121: Obtain the person image from the target camera through the video recorder and extract the person features of the person image.

[0079] S122: Perform feature matching on the person features in the personnel information database and determine the personnel identity of the personnel in the person image according to the matching result.

[0080] In this embodiment, the computer device can directly lock the target camera through the video recorder and obtain the person image collected by it to extract the person features of the person image; then the computer device can perform feature matching on the person features in the personnel information database and determine the personnel identity of the personnel in the person image according to the matching result.

[0081] Specifically, to reduce the consumption of computing resources, the computer device can be linked and controlled with a video recorder to directly locate and lock the device currently marked as the target camera, and extract the latest captured person image from the real-time image stream or cached image of the camera. Then, the video recorder can call the built-in feature extraction algorithm to analyze the features of the person in the person image, and extract the person features according to the analysis results. Since multiple person information is pre-registered in the person information database, and each person information includes a person identity and person features, the computer device can directly match and compare the person features with the person features of the registered persons in the person information database, and quickly screen out the corresponding matching results through methods such as similarity calculation or vector distance measurement, and then determine the person identity of the person in the person image according to the matching results.

[0082] Further, before extracting features from the person image, the computer device can also perform preprocessing operations on the person image, including but not limited to normalization processing, sharpening processing, denoising processing, etc. Among them, normalization processing refers to mapping the data of each dimension of the image data vector to an interval between (0, 1) or (-1, 1), or mapping a certain norm of the data vector to 1; sharpening processing refers to compensating the contour of the image, enhancing the edges and the parts with gray level jumps of the image, making the image clear, which can be divided into two categories: spatial domain processing and frequency domain processing, by highlighting the edges and contours of the ground objects in the image, or the features of some linear target elements, to improve the contrast between the edges of the ground objects and the surrounding pixels; denoising processing refers to the process of reducing noise in digital images. Generally, during the digitization and transmission of images, they are often affected by imaging device and external environmental noise interference, etc., that is, the received image information generally includes noise, and these noises will become an important reason for image interference. By performing denoising processing on the image, the noise in the image can be removed, and the authenticity and accuracy of the obtained image can be further improved.

[0083] In one embodiment, the process of determining the person identity of the person in the person image according to the matching result in step S122 may include:

[0084] S1221: Determine whether the matching result is empty.

[0085] S1222: If so, generate a temporary ID, bind the temporary ID to the person features and store them in the person information database, and use the temporary ID as the person identity of the person in the person image.

[0086] S1223: If not, directly obtain the identity ID in the matching result as the person identity of the person in the person image.

[0087] In this embodiment, after the computer device obtains the matching result, it can determine whether the matching result is empty, and then adopt different methods to determine the identity of the person according to the judgment result, so as to quickly access unregistered temporary personnel and improve the personnel management efficiency.

[0088] Specifically, if the matching result is empty, it means that there is no record in the personnel information database that matches the current personnel characteristics. At this time, the computer device can consider this person as an unregistered person or a temporary person who first enters the system recognition range, and immediately automatically generate a unique temporary identity identifier ID, and bind the temporary ID to the currently collected personnel characteristics. At the same time, the bound record is stored in the personnel information database as a new piece of data. In this way, even in the face of unfamiliar personnel, the computer device can achieve effective identity management and trajectory tracking without missing any key data.

[0089] In addition, if the matching result is not empty, it means that the computer device has successfully matched the registered identity information corresponding to the person image from the personnel information database. At this time, the computer device can directly read the personnel identity ID in the matching result and use it as the unique identity identifier of the person in the person image. All subsequent operations such as trajectory positioning update of this person will be based on this identity ID, so as to ensure the uniqueness of the personnel identity in the system and the continuity of data processing.

[0090] In one embodiment, the process of determining and updating the real-time trajectory positioning map of a person based on the person's identity in step S130 may include:

[0091] S131: Determine whether there is a real-time trajectory positioning map of the person's identity in the target area.

[0092] S132: If so, obtain the real-time trajectory positioning map and update the movement trajectory in the real-time trajectory positioning map according to the installation position of the target camera.

[0093] S133: If not, create a real-time trajectory positioning map of the person's identity in the target area and add the movement trajectory of the person to the real-time trajectory positioning map according to the installation position of the target camera.

[0094] In this embodiment, when updating the trajectory positioning of a person, the computer device may first determine whether there is a real-time trajectory positioning map of the person's identity in the target area; if not, it indicates that the person is detected in the target area for the first time. At this time, the computer device may create a real-time trajectory positioning map of the person's identity in the target area and add the person's movement trajectory to the real-time trajectory positioning map according to the installation position of the target camera; if it exists, it indicates that the person has left a movement trajectory in the target area. At this time, the computer device may directly obtain the corresponding real-time trajectory positioning map and update the movement trajectory in the real-time trajectory positioning map according to the installation position of the target camera.

[0095] Specifically, for the newly created real-time trajectory positioning map, the computer device may combine the installation position of the currently identified target camera and use the time, location, and related information of the person's appearance under this camera as the trajectory starting point and add it to the positioning map. For the existing real-time trajectory positioning map, the computer device may append the new movement trajectory points to the original path according to the installation position of the newly marked target camera to achieve continuous update of the trajectory; during the update process, the computer device may synchronously record information such as timestamps and position coordinates, and combine the connected path between the current target camera and the previous target camera to append the trajectory in the real-time trajectory positioning map to ensure that the entire trajectory has a complete time series and spatial coherence.

[0096] In one embodiment, the process of determining at least one camera path-connected to the target camera according to the topological relationship diagram and marking it as a candidate camera in step S130 may include:

[0097] S134: Determine the connected path corresponding to the target camera in the topological relationship diagram and mark the camera at the other end of each connected path as a candidate camera.

[0098] In this embodiment, after the person leaves the monitoring range of the target camera, the computer device may immediately perform path connectivity analysis on the position node where the current target camera is located based on the topological relationship diagram of the target area, and identify the connected path corresponding to the target camera, that is, all paths with effective path connections. Then, the computer device may traverse along the connected path and extract the camera nodes corresponding to the other end of each connected path. These cameras are regarded as the new position areas where the person is most likely to appear. Therefore, the computer device may mark them as candidate cameras and start a key monitoring mode for these candidate cameras to avoid the consumption of computing resources caused by blind search.

[0099] In one embodiment, the method may further include:

[0100] S150: After performing real-time monitoring on the candidate cameras, record the monitoring duration of the candidate cameras until a person appears within the monitoring range of a candidate camera is detected.

[0101] S160: If the monitoring duration exceeds a preset threshold, stop the real-time monitoring of the candidate cameras, and perform trajectory analysis on the real-time trajectory positioning map, and re-determine the candidate cameras based on the analysis results and then perform real-time monitoring.

[0102] In this embodiment, after the computer device performs real-time monitoring on the candidate cameras, it can record the monitoring duration of the candidate cameras until a person appears within the monitoring range of a candidate camera is detected; if the monitoring duration exceeds a preset threshold, stop the real-time monitoring of the candidate cameras, and perform trajectory analysis on the real-time trajectory positioning map, and re-determine the candidate cameras based on the analysis results and then perform real-time monitoring.

[0103] Specifically, during the real-time monitoring process, the computer device can simultaneously record in the background the cumulative monitoring duration from the start of monitoring to the current time point, and use it as an important parameter to measure the recognition delay and the target reappearance probability. When a certain candidate camera successfully identifies the target person within the preset time threshold, that is, completes the effective connection with the previous trajectory point, the computer device can mark this camera as the new target camera and update the real-time trajectory positioning map of the person. However, if the monitoring durations of all candidate cameras exceed this time threshold and the target person has still not been successfully detected, the computer device can determine that the current candidate camera group fails to cover the actual moving direction of the person. At this time, it can actively stop the real-time monitoring of these candidate cameras to save computing resources and avoid redundant processing.

[0104] Next, the computer device can immediately perform trajectory analysis on the existing real-time trajectory positioning map of this person. By analyzing data such as its historical movement pattern, path selection tendency, movement speed, and time interval, it can speculate on the possible deviation path behavior of the person, and accordingly re-evaluate the potential path nodes in the topological map, so as to select a new group of candidate cameras. These new candidate cameras will be included in the real-time monitoring range by the computer device again to form a new round of recognition loop until the target person is successfully re-identified, ensuring that the entire tracking process still has high adaptability and continuity in the case where the person deviates from the normal path or is not recognized in time, thereby effectively improving the accuracy and robustness of trajectory tracking.

[0105] The target trajectory positioning device provided by the embodiments of the present application will be described below. The target trajectory positioning device described below can be mutually referred to with the target trajectory positioning method described above.

[0106] In one embodiment, as Figure 3 shown, Figure 3Structural schematic diagram of a target trajectory positioning device provided by an embodiment of the present application; the present application also provides a target trajectory positioning device, including a personnel detection module 210, an identity determination module 220, a trajectory update module 230, and a cross-camera monitoring module 240, specifically including the following:

[0107] The personnel detection module 210 is configured to, after detecting that a person enters the target area, obtain a topological relationship diagram of cameras in the target area, and mark the camera whose monitoring range covers the person as the target camera; wherein, cameras are deployed at each path bifurcation in the target area.

[0108] The identity determination module 220 is configured to collect an image of the person through the target camera to obtain a person image, and determine the identity of the person in the person image based on the personnel information database.

[0109] The trajectory update module 230 is configured to determine and update the real-time trajectory positioning map of the person based on the personnel identity, and after the person leaves the monitoring range of the target camera, determine at least one camera that is path-connected to the target camera according to the topological relationship diagram, and mark it as the candidate camera.

[0110] The cross-camera monitoring module 240 is configured to perform real-time monitoring on the candidate cameras, and when it is detected that a person appears in the monitoring range of a candidate camera, mark the corresponding candidate camera as the new target camera, and return to collect an image of the person through the target camera to obtain a person image and its subsequent steps until it is detected that the person leaves the target area.

[0111] In the above embodiment, after detecting that a person enters the target area, a topological relationship diagram of cameras in the target area can be obtained, and the camera whose monitoring range covers the person can be marked as the target camera; wherein, cameras are only deployed at each path bifurcation in the target area, so that while reducing the hardware deployment cost, the comprehensive tracking of the person's trajectory can be ensured. Then, an image of the person can be collected through the target camera to obtain a person image, and the identity of the person in the person image can be determined based on the personnel information database. Then, the real-time trajectory positioning map of the person can be determined and updated based on the personnel identity, thereby improving the accuracy of continuous trajectory update. After the person leaves the monitoring range of the target camera, at least one camera that is path-connected to the target camera can be determined according to the topological relationship diagram and marked as the candidate camera, avoiding blindly searching all cameras and improving the positioning efficiency; finally, the candidate cameras can be monitored in real time, and when it is detected that a person appears in the monitoring range of a candidate camera, the corresponding candidate camera can be marked as the new target camera to continue to perform positioning update on the person's trajectory through the target camera, realizing continuous and seamless tracking of the person until the person leaves the target area, ensuring the integrity of trajectory positioning.

[0112] In one embodiment, the personnel detection module 210 may include:

[0113] A data acquisition sub-module, configured to acquire the path structure of the target area and the installation positions of each camera.

[0114] A matrix establishment sub-module, configured to analyze the physical connectivity between each installation position according to the path structure, and establish a reachability matrix between each camera according to the analysis result.

[0115] A relationship graph construction sub-module, configured to determine the connection paths between each adjacent and connected camera based on the reachability matrix, and construct a topological relationship graph according to each camera and its corresponding connection path.

[0116] In one embodiment, the identity determination module 220 may include:

[0117] A feature extraction sub-module, configured to obtain a person image from a target camera through a video recorder, and extract the person features of the person image.

[0118] A feature matching sub-module, configured to perform feature matching on the person features in the personnel information database, and determine the personnel identity of the person in the person image according to the matching result.

[0119] Wherein, a plurality of personnel information is pre-registered in the personnel information database, and each personnel information includes a personnel identity and personnel features.

[0120] In one embodiment, the feature matching sub-module may include:

[0121] A result judgment unit, configured to judge whether the matching result is empty.

[0122] A first identity determination unit, configured to generate a temporary ID when the matching result is empty, bind the temporary ID to the personnel features and store them in the personnel information database, and use the temporary ID as the personnel identity of the person in the person image.

[0123] A second identity determination unit, configured to directly obtain the identity ID in the matching result as the personnel identity of the person in the person image when the matching result is not empty.

[0124] In one embodiment, the trajectory update module 230 may include:

[0125] A positioning map judgment sub-module, configured to judge whether there is a real-time trajectory positioning map of the personnel identity in the target area.

[0126] The first trajectory update sub-module is used to obtain the real-time trajectory positioning map when the personnel identity has a real-time trajectory positioning map in the target area, and update the moving trajectory in the real-time trajectory positioning map according to the installation position of the target camera.

[0127] The second trajectory update sub-module is used to create a real-time trajectory positioning map of the personnel identity in the target area when there is no real-time trajectory positioning map of the personnel identity in the target area, and add the moving trajectory of the personnel to the real-time trajectory positioning map according to the installation position of the target camera.

[0128] In one embodiment, the trajectory update module 230 may further include:

[0129] The camera candidate sub-module is used to determine the connected path corresponding to the target camera in the topological relationship graph, and mark the camera at the other end of each connected path as a candidate camera.

[0130] In one embodiment, the device may further include:

[0131] The duration monitoring module is used to record the monitoring duration of the candidate camera after real-time monitoring of the candidate camera until a person appears in the monitoring range of the candidate camera.

[0132] The monitoring update module is used to stop the real-time monitoring of the candidate camera if the monitoring duration exceeds the preset threshold, and perform trajectory analysis on the real-time trajectory positioning map, and re-determine the candidate camera according to the analysis result and perform real-time monitoring.

[0133] In one embodiment, the present application also provides a storage medium, in which computer-readable instructions are stored. When the computer-readable instructions are executed by one or more processors, the one or more processors are caused to execute the steps of the target trajectory positioning method as described in any one of the above embodiments.

[0134] In one embodiment, the present application also provides a computer device, in which computer-readable instructions are stored. When the computer-readable instructions are executed by one or more processors, the one or more processors are caused to execute the steps of the target trajectory positioning method as described in any one of the above embodiments.

[0135] Schematically, as Figure 4 shown, Figure 4 is an internal structural schematic diagram of a computer device provided by an embodiment of the present application. The computer device 300 may be provided as a server. Refer to Figure 4, the computer device 300 includes a processing component 302, which further includes one or more processors, and memory resources represented by a memory 301 for storing instructions executable by the processing component 302, such as application programs. The application programs stored in the memory 301 may include one or more modules each corresponding to a set of instructions. In addition, the processing component 302 is configured to execute instructions to perform the target trajectory positioning method of any of the above embodiments.

[0136] The computer device 300 may further include a power supply component 303 configured to perform power management of the computer device 300, a wired or wireless network interface 304 configured to connect the computer device 300 to a network, and an input / output (I / O) interface 305. The computer device 300 may operate based on an operating system stored in the memory 301, such as Windows Server TM, Mac OS XTM, Unix TM, Linux TM, Free BSDTM or the like.

[0137] Those skilled in the art can understand that Figure 4 the structure shown in

[0138] is only a block diagram of a part of the structure related to the solution of this application, and does not constitute a limitation on the computer device to which the solution of this application is applied. A specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different component layout.

[0139] Each embodiment in this specification is described in a progressive manner. The key point of each embodiment is to illustrate the differences from other embodiments. The embodiments can be combined as needed, and the same or similar parts can be referred to each other.

[0140] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but rather to the broadest scope consistent with the principles and novel features disclosed herein.

Claims

1. A target trajectory positioning method, characterized in that, The method includes: After detecting that a person enters the target area, obtain the topological relationship diagram of the cameras in the target area, and mark the cameras whose monitoring ranges cover the person as target cameras; wherein, cameras are deployed at each path bifurcation in the target area; Collect an image of the person through the target camera to obtain a person image, and determine the identity of the person in the person image based on the personnel information database; Determine and update the real-time trajectory positioning map of the person based on the identity of the person, and after the person leaves the monitoring range of the target camera, determine at least one camera that is path-connected to the target camera according to the topological relationship diagram, and mark it as a candidate camera; Monitor the candidate cameras in real time, and when it is detected that the person appears in the monitoring range of a candidate camera, mark the corresponding candidate camera as the new target camera, and return to collect an image of the person through the target camera to obtain a person image and its subsequent steps until it is detected that the person leaves the target area.

2. The target trajectory positioning method according to claim 1, characterized in that The obtaining of the topological relationship diagram of the cameras in the target area includes: Obtain the path structure of the target area and the installation positions of each camera; Analyze the physical connectivity between each installation position according to the path structure, and establish an accessibility matrix between each camera according to the analysis result; Determine the connectivity paths between each adjacent and connected camera based on the accessibility matrix, and construct a topological relationship diagram according to each camera and its corresponding connectivity path.

3. The target trajectory positioning method according to claim 1, characterized in that, The determining of the identity of the person in the person image based on the personnel information database includes: Obtain the person image from the target camera through a video recorder, and extract the person features of the person image; Perform feature matching on the person features in the personnel information database, and determine the identity of the person in the person image according to the matching result; Wherein, multiple personnel information are registered in advance in the personnel information database, and each personnel information includes an identity and person features.

4. The target trajectory positioning method according to claim 3, characterized in that, The determining of the identity of the person in the person image according to the matching result includes: Judge whether the matching result is empty; If so, generate a temporary ID, bind the temporary ID to the person features and store them in the personnel information database, and use the temporary ID as the identity of the person in the person image; If not, directly obtain the identity ID in the matching result as the identity of the person in the person image.

5. The target trajectory positioning method according to claim 1, wherein The determining and updating of the real-time trajectory positioning map of the person based on the identity of the person includes: Judge whether there is a real-time trajectory positioning map of the person identity in the target area; If so, obtain the real-time trajectory positioning map, and update the movement trajectory in the real-time trajectory positioning map according to the installation position of the target camera; If not, create a real-time trajectory positioning map of the person identity in the target area, and add the movement trajectory of the person to the real-time trajectory positioning map according to the installation position of the target camera.

6. The target trajectory positioning method according to claim 1, wherein Determining at least one camera that is connected to the target camera path according to the topological relationship diagram and marking it as a candidate camera includes: Determining the connected path corresponding to the target camera in the topological relationship diagram and marking the camera at the other end of each connected path as a candidate camera.

7. The target trajectory positioning method according to claim 1, characterized in that, The method further includes: After real-time monitoring of the candidate cameras, recording the monitoring duration of the candidate cameras until it is detected that the person appears in the monitoring range of a candidate camera; If the monitoring duration exceeds a preset threshold, stop the real-time monitoring of the candidate cameras, perform trajectory analysis on the real-time trajectory positioning map, and re-determine candidate cameras according to the analysis results and perform real-time monitoring.

8. A target trajectory positioning device, characterized in that, It includes: A personnel detection module, configured to, after detecting that a person enters the target area, obtain the topological relationship diagram of the cameras in the target area, and mark the camera whose monitoring range covers the person as the target camera; wherein, cameras are deployed at each path bifurcation in the target area; An identity determination module, configured to collect an image of the person through the target camera to obtain a person image, and determine the identity of the person in the person image based on the personnel information database; A trajectory update module, configured to determine and update the real-time trajectory positioning map of the person based on the personnel identity, and after the person leaves the monitoring range of the target camera, determine at least one camera that is connected to the target camera path according to the topological relationship diagram and mark it as a candidate camera; A cross-camera monitoring module, configured to perform real-time monitoring on the candidate cameras, and when it is detected that the person appears in the monitoring range of a candidate camera, mark the corresponding candidate camera as the new target camera, and return to collect an image of the person through the target camera to obtain a person image and its subsequent steps until it is detected that the person leaves the target area.

9. A storage medium, characterized in that: The computer-readable instructions are stored in the storage medium, and when the computer-readable instructions are executed by one or more processors, the one or more processors are caused to execute the steps of the target trajectory positioning method according to any one of claims 1 to 7.

10. A computer device, characterized in that, It includes: One or more processors and a memory; The computer-readable instructions are stored in the memory, and when the computer-readable instructions are executed by the one or more processors, the steps of the target trajectory positioning method according to any one of claims 1 to 7 are executed.

Images