Target trajectory tracking method and device, server and storage medium

Abstract

The invention relates to the technical field of computer vision, in particular to a target trajectory tracking method and device, a server and a storage medium, and is used for solving the problem that the position of a target in an actual scene cannot be positioned and displayed in related technologies. The method comprises the following steps: based on feature information obtained by performing target detection on acquired image data containing a to-be-tracked target, performing at least one time of retrieval in a plurality of pieces of detection data to obtain each piece of detection data associated with the feature information, any detection data comprises identification information of a target, feature information of the target in a basic image associated with the detection data, identification information of each associated target, acquisition time of the basic image and a geographic position of an acquisition device; and according to the acquisition time and the geographic position included in each detection data, determining the spatial-temporal change information of the to-be-tracked target, thereby drawing and displaying the movement trajectory of the to-be-tracked target based on the spatial-temporal change information, and realizing macroscopic movement trajectory tracking of the target.

Description

Technical Field

[0001] This application relates to the field of computer vision technology, and in particular, to a method, apparatus, server, and storage medium for target trajectory tracking. Background Art

[0002] With the development of the field of computer vision, the application of target movement trajectory analysis has become increasingly widespread. In related technologies, most target trajectory tracking methods are performed on the RGB image information of a single video or video file. This method can determine that the target appears in the area monitored by the camera and can accurately determine that the target moves from one point to another within this area, so that the accurate movement trajectory of the target can be drawn within the video frame; however, it cannot locate and display the position of the target in the actual scene, nor can it accurately depict the movement trajectory of the target in the actual scene.

[0003] Currently, in many fields, such as intelligent security, smart communities, and smart buildings, the demand for obtaining and displaying the position of the target in the actual scene is increasing, and the above-mentioned target trajectory tracking method can no longer meet this demand. Summary of the Invention

[0004] Embodiments of this application provide a method, apparatus, server, and storage medium for target trajectory tracking, which can perform macroscopic movement trajectory tracking on the target, and can relatively clearly obtain the movement trajectory of the target, as well as the activity range path, etc.

[0005] The specific technical solutions provided by the embodiments of this application are as follows:

[0006] In a first aspect, an embodiment of this application provides a method for target trajectory tracking, including:

[0007] Obtain image data including the target to be tracked, perform target detection on the image data, and obtain the feature information of the target to be tracked;

[0008] Based on the feature information, perform at least one retrieval among multiple detection data to obtain each detection data associated with the feature information, where any one of the multiple detection data includes the identification information of the target, the feature information of the target in the base image associated with the any one of the detection data, and the identification information of each associated target, as well as the acquisition time of the base image and the geographical location of the acquisition device;

[0009] Determine the spatio-temporal change information of the target to be tracked according to the acquisition time and geographical location included in each detection data;

[0010] Based on the spatio-temporal change information, draw and display the movement trajectory of the target to be tracked.

[0011] In a possible implementation method, retrieving at least once among multiple detection data based on the feature information to obtain each detection data associated with the feature information includes:

[0012] Based on the feature information, obtaining the detection data of the target that matches the feature information from the multiple detection data, and the detection data corresponding to the identification information of each target included in the detection data;

[0013] When it is determined that the retrieval times threshold is not reached, repeat the following operations until the retrieval times threshold is reached:

[0014] For any retrieval, taking each currently obtained target as a retrieval object, and for any retrieval object, based on the feature information of the any retrieval object, obtaining the detection data of the candidate target that matches the feature information of the any retrieval object from the multiple detection data, and

[0015] the detection data corresponding to the identification information of each target included in the detection data of the candidate target;

[0016] Taking each obtained detection data as the detection data associated with the feature information to obtain the each detection data.

[0017] In a possible implementation method, determining the spatio-temporal change information of the target to be tracked according to the acquisition time and geographical location included in the each detection data includes:

[0018] Adding the retrieval cumulative times and feature matching degrees when the each detection data is obtained to the respective detection data to obtain each retrieval result data;

[0019] According to the preset geographical blocks, dividing the retrieval result data located in the same geographical block among the each retrieval result data into the same data set to obtain multiple data sets;

[0020] Based on the acquisition time, geographical location, retrieval cumulative times and feature matching degrees included in the retrieval result data in the multiple data sets, respectively determining the confidence degrees of the acquisition time and geographical location pair of the target to be tracked in each data set;

[0021] Taking the acquisition time and geographical location pair corresponding to the confidence degrees greater than or equal to the confidence threshold among the determined each confidence degree as the spatio-temporal change information of the target to be tracked.

[0022] In a possible implementation method, by performing the following operations, determining the confidence degrees of the acquisition time and geographical location pair of the target to be tracked in each data set:

[0023] Determine the splitting index value of any dataset based on the average acquisition time of the retrieval result data in the dataset and the standard deviation of the acquisition time.

[0024] If it is determined that the splitting index value is greater than a preset coefficient of variation constant, then based on the acquisition time, geographical location, cumulative number of retrievals, and feature matching degree included in the retrieval result data in any dataset, determine the confidence of the acquisition time and geographical location pair of the target to be tracked in any dataset.

[0025] If it is determined that the splitting index value is less than or equal to the coefficient of variation constant, then based on the coefficient of variation constant, split any dataset in the time dimension so that the splitting index values corresponding to the split sub-datasets are greater than the coefficient of variation constant; based on the acquisition time, geographical location, cumulative number of retrievals, and feature matching degree included in the retrieval result data in each sub-dataset, respectively determine the confidence of the acquisition time and geographical location pair of the target to be tracked in each sub-dataset.

[0026] In a possible implementation method, the drawing and display of the movement track of the target to be tracked based on the spatio-temporal change information include:

[0027] Based on the map data, map each geographical location in the spatio-temporal change information to the corresponding map in the map data to obtain each trajectory point.

[0028] Connect each trajectory point in sequence according to each acquisition time in the spatio-temporal change information to obtain the movement track of the target to be tracked.

[0029] Display the movement track in a preset interface.

[0030] In a possible implementation method, after displaying the movement track, the method further includes:

[0031] In response to a display operation for candidate detection data associated with any trajectory point in the movement track, display the candidate detection data and the associated candidate base image.

[0032] If an additional retrieval instruction for an associated target in the candidate base image is received, then based on the additional retrieval instruction, retrieve the detection data associated with the associated target from the multiple detection data, where the additional retrieval instruction includes the associated target, which is a target strongly correlated with the target to be tracked included in the image data.

[0033] Add the retrieved detection data to the detection data, and redraw the movement track of the target to be tracked based on the newly obtained detection data.

[0034] In a possible implementation method, retrieving, from the multiple detection data, the detection data associated with the associated target based on the additional retrieval instruction includes:

[0035] Obtaining, from the multiple detection data, the detection data of a reference target that matches the feature information of the associated target based on the feature information of the associated target;

[0036] Obtaining, from the multiple detection data, the detection data corresponding to the identification information of each target included in the detection data of the reference target;

[0037] Determining each obtained detection data as the detection data associated with the associated target respectively.

[0038] In a possible implementation method, if the identification information of the acquisition device carried by the image data is the identification information of a registered acquisition device, the method further includes:

[0039] Obtaining the device geographical location corresponding to the identification information of the acquisition device carried by the image data;

[0040] Connecting the last trajectory point in the movement trajectory to the trajectory point of the device geographical location on the map, and obtaining and displaying the final movement trajectory of the target to be tracked.

[0041] In a second aspect, an embodiment of the present application provides a target trajectory tracking device, including:

[0042] A target detection module, configured to obtain image data including a target to be tracked, perform target detection on the image data, and obtain the feature information of the target to be tracked;

[0043] A retrieval module, configured to perform at least one retrieval in multiple detection data based on the feature information, and obtain each detection data associated with the feature information, where any detection data in the multiple detection data includes the identification information of a target, the feature information of the target in a base image associated with the any detection data and the identification information of associated targets, and the acquisition time of the base image and the geographical location of the acquisition device;

[0044] A data aggregation module, configured to determine the spatio-temporal change information of the target to be tracked according to the acquisition time and geographical location included in each detection data;

[0045] A display module, configured to draw and display the movement trajectory of the target to be tracked based on the spatio-temporal change information.

[0046] In a possible implementation manner, the retrieval module is specifically configured to:

[0047] Based on the feature information, obtain, from the multiple detection data, the detection data of the target that matches the feature information, and the detection data corresponding to the identification information of each target included in the detection data;

[0048] When it is determined that the retrieval count threshold has not been reached, repeat the following operations until the retrieval count threshold is reached:

[0049] For any one retrieval, use each currently obtained target as a retrieval object. For any one retrieval object, based on the feature information of the any one retrieval object, obtain, from the multiple detection data, the detection data of the candidate target that matches the feature information of the any one retrieval object, and

[0050] the detection data corresponding to the identification information of each target included in the detection data of the candidate target;

[0051] Use each obtained detection data as the detection data associated with the feature information to obtain the respective detection data.

[0052] In a possible implementation manner, the data aggregation module is specifically configured to:

[0053] Add the retrieval cumulative count and the feature matching degree when each detection data is obtained to the respective detection data to obtain each retrieval result data;

[0054] According to the preset geographical regions, divide the retrieval result data located in the same geographical region among the respective retrieval result data into the same data set to obtain multiple data sets;

[0055] Based on the collection time, geographical location, retrieval cumulative count, and feature matching degree included in the retrieval result data in the multiple data sets, determine the confidence degree of the collection time and geographical location pair of the target to be tracked in each data set respectively;

[0056] Use the collection time and geographical location pair corresponding to the confidence degree greater than or equal to the confidence threshold among the determined confidence degrees as the spatio-temporal change information of the target to be tracked.

[0057] In a possible implementation manner, by performing the following operations, determine the confidence degree of the collection time and geographical location pair of the target to be tracked in each data set:

[0058] Based on the average collection time of the retrieval result data in any one data set and the standard deviation of the collection time, determine the splitting index value of the any one data set;

[0059] If it is determined that the split index value is greater than a preset coefficient of variation constant, then based on the collection time, geographical location, cumulative retrieval times, and feature matching degree included in the retrieval result data in any one of the data sets, determine the confidence level of the pair of collection time and geographical location of the target to be tracked in any one of the data sets;

[0060] If it is determined that the split index value is less than or equal to the coefficient of variation constant, then based on the coefficient of variation constant, split any one of the data sets in the time dimension so that the split index values corresponding to the split sub-data sets are greater than the coefficient of variation constant; based on the collection time, geographical location, cumulative retrieval times, and feature matching degree included in the retrieval result data in each sub-data set, determine the confidence level of the pair of collection time and geographical location of the target to be tracked in each sub-data set respectively.

[0061] In a possible implementation manner, the display module is specifically configured to:

[0062] Based on the map data, map each geographical location in the spatio-temporal change information to the corresponding map in the map data respectively to obtain each trajectory point;

[0063] Connect the respective trajectory points in sequence according to each collection time in the spatio-temporal change information to obtain the movement trajectory of the target to be tracked;

[0064] Display the movement trajectory in a preset interface.

[0065] In a possible implementation manner, after displaying the movement trajectory, the display module is further configured to:

[0066] In response to a display operation on candidate detection data associated with any one of the trajectory points in the movement trajectory, display the candidate detection data and the associated candidate base image;

[0067] If an additional retrieval instruction for an associated target in the candidate base image is received, then based on the additional retrieval instruction, retrieve the detection data associated with the associated target from the multiple detection data, where the additional retrieval instruction includes the associated target, which is a target strongly correlated with the target to be tracked included in the image data;

[0068] Add the retrieved detection data to the respective detection data, and redraw the movement trajectory of the target to be tracked based on the newly obtained respective detection data.

[0069] In a possible implementation manner, the display module is specifically configured to:

[0070] Based on the feature information of the associated target, obtain the detection data of the reference target that matches the feature information of the associated target from the multiple detection data;

[0071] From the multiple detection data, obtain the detection data corresponding to the identification information of each target included in the detection data of the reference target;

[0072] Determine each obtained detection data as the detection data associated with the associated target respectively.

[0073] In a possible implementation manner, if the identification information of the acquisition device carried by the image data is the identification information of a registered acquisition device, the display module is further configured to:

[0074] Obtain the device geographical location corresponding to the identification information of the acquisition device carried by the image data;

[0075] Connect the last track point in the movement track to the track point of the device geographical location on the map, and obtain and display the final movement track of the target to be tracked.

[0076] In a third aspect, an embodiment of the present application provides a server, including:

[0077] A memory, configured to store computer programs or instructions;

[0078] A processor, configured to execute the computer programs or instructions in the memory, so that the method according to any one of the first aspects described above is executed.

[0079] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, when the instructions in the storage medium are executed by a processor, enabling the processor to execute the method according to any one of the first aspects described above.

[0080] In a fifth aspect, an embodiment of the present application provides a computer program product, the computer program product includes: computer program code, when the computer program code runs on a computer, enabling the computer to execute the method according to any one of the first aspects described above.

[0081] In the embodiments of the present application, image data including a target to be tracked is acquired, and target detection is performed on the image data to obtain feature information of the target to be tracked; based on the feature information, at least one retrieval is performed among multiple detection data to obtain each detection data associated with the feature information, where any one of the multiple detection data includes identification information of the target, feature information of the target in the base image associated with the detection data, identification information of each associated target, as well as the acquisition time of the base image and the geographical location of the acquisition device; according to the acquisition time and geographical location included in each detection data, the spatio-temporal change information of the target to be tracked is determined; in this way, based on the determined spatio-temporal change information, the movement trajectory of the target to be tracked can be drawn and then displayed. Since the spatio-temporal change information includes real time and real geographical location, the displayed movement trajectory realizes macroscopic movement trajectory tracking of the target, so that the movement trajectory and activity range path of the target can be obtained more clearly. Description of the Drawings

[0082] Figure 1 It is a schematic diagram of an application scenario of an optional target trajectory tracking method in the embodiments of the present application;

[0083] Figure 2 It is a schematic flowchart of a target trajectory tracking method in the embodiments of the present application;

[0084] Figure 3 It is a schematic diagram of an image of a base image in the embodiments of the present application;

[0085] Figure 4 It is a schematic flowchart of a specific retrieval process for obtaining each detection data in the embodiments of the present application;

[0086] Figure 5 It is a schematic flowchart of a process for determining the spatio-temporal change information of the target to be tracked in the embodiments of the present application;

[0087] Figure 6 It is a schematic flowchart of another process for determining the spatio-temporal change information of the target to be tracked in the embodiments of the present application;

[0088] Figure 7 It is a schematic flowchart of a process for drawing a movement trajectory in the embodiments of the present application;

[0089] Figure 8 It is a schematic diagram of an optional simple road scene in the embodiments of the present application;

[0090] Figure 9 It is a schematic diagram of the segmentation effect of an optional road scene in the embodiments of the present application;

[0091] Figure 10It is an optional schematic diagram of a moving trajectory drawn based on screening points in an embodiment of the present application;

[0092] Figure 11 It is an optional schematic diagram of a moving trajectory drawn based on additional retrieval targets in an embodiment of the present application;

[0093] Figure 12 It is a schematic diagram of the logical architecture of a target trajectory tracking device in an embodiment of the present application;

[0094] Figure 13 It is a schematic diagram of the entity architecture of an optional server in an embodiment of the present application. Detailed implementation manners

[0095] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

[0096] It should be noted that the term "and / or" in the specification, claims and drawings of the present application describes the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist simultaneously, and B exists alone. The character " / " generally represents an "or" relationship between the associated objects before and after.

[0097] The terms "first", "second", "third", etc. in the specification, claims and the above-mentioned drawings of the present application are used to distinguish similar objects, and do not have to be used to describe a specific order or sequence. It should be understood that such data can be interchanged under appropriate circumstances so that the embodiments of the present application described here can be implemented in an order other than those illustrated or described here.

[0098] In the technical solutions of the present application, the collection, dissemination, use, etc. of data all comply with the requirements of relevant national laws and regulations.

[0099] Before introducing the target trajectory tracking method provided by the embodiments of the present application, for the convenience of understanding, the technical background of the embodiments of the present application will be first introduced in detail below.

[0100] Currently, in many fields, such as intelligent security, smart communities, and intelligent buildings, the demand for obtaining the position of a target in an actual scenario and displaying it is increasing. However, in related technologies, most target trajectory tracking methods are based on the RGB image information of a single video or video file. This method can determine that the target appears in the area monitored by the camera and can accurately determine that the target moves from one point to another within this area, so that the accurate movement trajectory of the target can be drawn within the video frame; but it cannot locate and display the position of the target in the actual scenario, nor can it accurately depict the movement trajectory of the target in the actual scenario.

[0101] In view of this, in order to solve the problem in related technologies that the position of a target in an actual scenario cannot be located and displayed, an embodiment of the present application provides a target trajectory tracking method. In the embodiment of the present application, image data containing the target to be tracked is obtained, and target detection is performed on the image data to obtain the feature information of the target to be tracked; based on the feature information, at least one retrieval is performed among multiple detection data to obtain each detection data associated with the feature information, where any one of the multiple detection data includes the identification information of the target, the feature information of the target in the base image associated with the detection data, and the identification information of each associated target, as well as the acquisition time of the base image and the geographical location of the acquisition device; according to the acquisition time and geographical location included in each detection data, the spatio-temporal change information of the target to be tracked is determined; in this way, based on the determined spatio-temporal change information, the movement trajectory of the target to be tracked can be drawn, and then the movement trajectory can be displayed.

[0102] Since the displayed movement trajectory of the target to be tracked is drawn based on the determined spatio-temporal change information, and the spatio-temporal change information includes the time and space information of the target to be tracked, that is, the geographical location where the target to be tracked is located at a certain time, and the spatio-temporal change information of the target to be tracked is determined based on the acquisition time of the base image and the geographical location of the acquisition device included in each detection data associated with the feature information of the target to be tracked, therefore, each trajectory point in the displayed movement trajectory can represent that the target to be tracked is located at a certain real geographical location at a certain real time, thus realizing the macroscopic movement trajectory tracking of the target, and the movement trajectory and activity range path of the target can be obtained more clearly.

[0103] Figure 1 It is a schematic diagram of the application scenario of an optional target trajectory tracking method in the embodiment of the present application. Refer to Figure 1As shown in the figure, the application scenario can be a traffic scenario. The target trajectory tracking system in this application scenario includes cameras installed in each block and a central data node. Among them, the cameras are used to collect block images and transmit the collected block images to the central data node. The central data node is used to execute a target trajectory tracking method in an embodiment of the present application, and display the moving trajectory of the target to be tracked through the front-end interface of the central data node.

[0104] In some embodiments, the cameras and the central data node can be connected by wired or wireless means. Optionally, Figure 1 As shown in the figure, the cameras and the central data node are connected by wireless means. It should be noted that the connection method of the present application is not limited.

[0105] In some embodiments, the central data node can be a server, such as an independent physical server, a server cluster or a distributed system composed of multiple physical servers, providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, Content Delivery Network (CDN), and big data and artificial intelligence platforms.

[0106] It should be noted that Figure 1 This is an example introduction to the application scenario of a target trajectory tracking method provided by an embodiment of the present application. In fact, the application scenarios applicable to the method in an embodiment of the present application are not limited to this, such as any scenario that needs to track the target trajectory in the fields of intelligent security, smart community, and smart building.

[0107] After introducing the application scenario of the embodiment of the present application, the following further details the preferred embodiments of the present application with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present application, and are not used to limit the present application. And without conflict, the embodiments of the present application and the features in the embodiments can be combined with each other.

[0108] Refer to Figure 2 As shown in the figure, an embodiment of the present application provides a target trajectory tracking method. The specific process of this method may include but is not limited to the following:

[0109] Step 200: Obtain image data containing the target to be tracked, perform target detection on the image data, and obtain the feature information of the target to be tracked.

[0110] In some embodiments, such as Figure 1 In the application scenario shown in the figure, when executing step 200, image data transmitted by a camera located at any geographical location can be received, where the image data includes the target to be tracked, such as denoted as TG ; The image data can be the monitored image data captured by a camera, or a frame of image data in the monitored video collected by a camera. The present application does not make specific limitations.

[0111] After receiving the image data, perform object detection on the image data to extract the target T to be tracked G 's feature information a. It should be noted that the present application does not limit the object detection method. Traditional object detection methods can be used to perform object detection on the image data, such as YOLO series algorithms, etc., to extract the feature information a of the target T to be tracked in the image data. G 's feature information a.

[0112] Step 210: Based on the feature information, perform at least one retrieval among multiple detection data to obtain each detection data associated with the feature information. Among them, any one of the multiple detection data includes the identification information of the target, the feature information of the target in the base image associated with the detection data, and the identification information of each associated target, as well as the acquisition time of the base image and the geographical location of the acquisition device.

[0113] In the embodiments of the present application, a certain number of base images are collected in advance, object detection and feature association are performed on the collected base images to obtain the foregoing multiple detection data, and the multiple detection data are stored in the central data node. Among them, the collected base images include, but are not limited to, images in videos, recordings, real-time monitored images, captured images, etc. collected by cameras distributed in each block.

[0114] Specifically, perform object detection on each collected base image to obtain the information of the target T in each base image, as well as the feature information of the target T, and determine the association relationship between the target T in each base image and other targets during the detection process.

[0115] As Figure 3 shown, assume that the currently processed base image contains two people. Among them, one person is riding a bicycle and wearing a helmet, and the other person is standing by the roadside with a single-shoulder bag on the back. Then, when performing object detection on this base image, the person riding the bicycle is detected and recorded as the target T. Since the target T is riding a bicycle and wearing a helmet, the targets associated with the target T are the bicycle and the helmet. Then, the association relationship between the target T in this base image and other targets can be expressed as: {T1, T2} ∈ T, where T1 represents the bicycle, T2 represents the helmet, and "2" represents the quantity, that is, the quantity of each associated target.

[0116] In some embodiments, each target may further include sub-targets. Among them, only those that can frame a region in the image can become sub-targets of the target.

[0117] Continuing with the foregoing examples, assuming the targets include people, bicycles, cars, etc. For people, their sub-targets can be divided according to clothing and can include upper garments, trousers, skirts, shoes, etc., or can be divided according to parts of the body and can include the head, upper body, lower body, feet, etc. Among them, the attributes of people can include gender, etc., and the attributes of upper garments / trousers / skirts can include color, style, etc.; for cars, their sub-targets can include wheels, car doors, license plates, the front of the car, etc.

[0118] For example, for the trademark of the helmet T2, etc., it can be used as a sub-target of the helmet T2. The association relationships between these targets can be expressed as {T 21} ∈ T2, where T 21 can represent the trademark of the helmet.

[0119] Correspondingly, referring to Figure 3 as shown, for a person standing by the roadside with a single-shoulder bag on the back, assuming it is denoted as the target O (objective), then the association relationship between the target O and its sub-targets in the obtained basic image can be expressed as: {O1, O2, O3, O4, O5} ∈ O, where O1 represents the bag carried; O2 represents the upper garment, and the attributes of the upper garment can include color, style, etc.; O3 represents the skirt, and the attributes of the skirt can also include color, style, etc.; O4 represents the shoes, and the attributes of the shoes can include color, type, style, etc.; O5 represents the head, and the attributes of the head can include hairstyle, hair color, etc.

[0120] It can be understood that the hierarchical depth of the association between targets, that is, {T1, T2} ∈ T, {T 21} ∈ T2, {T 211 , …} ∈ T 21 , …… the vertical extension depth can be set according to actual needs. In some embodiments, it can only include the first-level association relationship of the target T such as people, cars, bicycles, etc.; in other embodiments, it can also include the sub-targets of the foregoing target T, and even more levels of association relationships such as the sub-targets of the sub-targets.

[0121] Optionally, in the embodiments of the present application, the two-level association relationship of the target T and the sub-targets of the target T can be selected to ensure the retrieval efficiency and performance consumption.

[0122] After completing object detection and feature association for the basic data (i.e., obtaining the association relationships between objects), structured data information for each object in the basic data is generated, which is any of the aforementioned detection data. Each piece of structured data information includes, but is not limited to, the identification information of the object (such as the unique number i), the acquisition time t of the associated basic image, the geographical location s of the acquisition device (such as a camera) of the associated basic image, the attribute c of the object, the feature information a of the object, the identification information of each object associated with the object in the basic image, etc. Among them, the identification information of each associated object can be stored in the form of a list, which can be denoted as T i , i ∈ [0, T iC .

[0123] In the embodiments of the present application, any detection data can be expressed as T(i, t, s, c, a, T iC ). It should be noted that the aforementioned storage method of the detection data can be row storage, column storage, or any other storage method, and the present application does not make specific limitations or restrictions; correspondingly, the storage medium of the detection data is also not specifically limited or restricted by the present application, and any storage medium such as a disk, a solid-state drive, or a memory can be used.

[0124] It should be noted that the aforementioned data collection process can be to collect data for a historical period of time or to collect real-time continuously updated data, and object detection and feature association are performed on the collected data for subsequent retrieval use.

[0125] In the present application, through the above method, object detection and feature association are performed on each pre-collected basic image, thereby obtaining a plurality of detection data, and these detection data are used for subsequent retrieval. The following will specifically describe the specific usage process in the retrieval stage in detail.

[0126] In some embodiments of the present application, as shown in Figure 4 , when performing step 210, the following steps can be specifically executed to obtain the detection data of each object associated with the feature information a:

[0127] Step 2101: Based on the feature information, obtain the detection data of the object that matches the feature information and the detection data corresponding to the identification information of each object included in the detection data from the plurality of detection data.

[0128] In the embodiments of the present application, since any one of the multiple detection data includes the identification information of the target, the feature information of the target in the base image associated with the detection data, and the identification information of each associated target, as well as the acquisition time of the base image and the geographical location of the acquisition device; then, in specific implementation, when performing step 2101, the feature information a of the target to be tracked can be respectively feature-matched with the feature information included in each of the multiple detection data to determine which or which of the multiple detection data the feature information of the target to be tracked matches, so as to retrieve from the multiple detection data the detection data that matches the feature information a of the target to be tracked. The target included in this detection data is a target similar to the target to be tracked. For the convenience of subsequent description, the target can be denoted as a similar target. Wherein, the matching of two pieces of feature information indicates that the matching rate (or called feature matching degree) between these two pieces of feature information is greater than the matching threshold.

[0129] After obtaining the detection data of the target that matches the feature information from the multiple detection data through the aforementioned feature matching, that is, the detection data of the similar target, according to the identification information of each target included in the obtained detection data, the detection data corresponding to the identification information of each target is obtained from the multiple detection data, where each target can be denoted as an associated target of the target to be tracked.

[0130] Step 2102: Determine whether the retrieval times threshold is reached; if not, then execute step 2103, if so, then execute step 2104.

[0131] In the embodiments of the present application, the retrieval times threshold can be preset, and the retrieval times threshold can be preset according to actual requirements. Optionally, in some embodiments of the present application, the retrieval times threshold can be selected as any value from 3 to 5.

[0132] In specific implementation, after performing step 2101, step 2102 is executed to determine whether the current cumulative retrieval times reaches the retrieval times threshold. If so, then step 2104 is executed. If not, then step 2103 is executed to further retrieve from the multiple detection data.

[0133] Step 2103: Repeat the following operations until the retrieval times threshold is reached: For any retrieval, each target obtained currently is used as a retrieval object. For any retrieval object, based on the feature information of the retrieval object, the detection data of the candidate target that matches the feature information of the retrieval object is obtained from the multiple detection data, and the detection data corresponding to the identification information of each target included in the detection data of the candidate target is obtained.

[0134] In the embodiments of the present application, when performing step 2103, the following operations are repeatedly performed until the retrieval times threshold is reached: In each retrieval, specifically, each currently obtained target is used as a retrieval object. For any retrieval object, the feature information of the retrieval object is feature-matched with the feature information of the targets included in multiple detection data, and the targets in the multiple detection data with a feature matching degree greater than the matching threshold are determined as candidate targets, that is, the similar targets similar to the retrieval object described above. Then, the detection data of the candidate target is screened out from the multiple detection data. Next, based on the identification information of each target included in the detection data of the candidate target, the detection data corresponding to the identification information of each target included in the detection data of the candidate target is obtained from the multiple detection data.

[0135] In some embodiments of the present application, when the retrieval depth (cumulative detection times) is 1, that is, after the first retrieval, assuming that there are m similar targets of the target to be tracked obtained by the retrieval, then, all the similar targets of the target to be tracked can be represented in a list form as {T1, T2,..., T m} ∈ T j , where j ∈ [1, m], and m is the serial number; all the associated targets included can be represented in a list form as T jq , where j ∈ [1, m], q ∈ [1, T jC , and the number of T jq is denoted as N1.

[0136] Then, the detection data obtained by the retrieval can be represented as Denoted as where "G" represents the target to be tracked, and "1" represents the retrieval depth, that is, the cumulative detection times.

[0137] If step 2102 is executed and it is determined that the retrieval times threshold is not reached, that is, a depth retrieval needs to be performed, then step 2103 is executed. All the associated targets in T jq are used as new retrieval objects in sequence, denoted as T G1 , and a retrieval is performed again to obtain the detection data of all the associated targets, denoted as S G1 .

[0138] When the retrieval depth (i.e., the cumulative retrieval times) is k, assuming that k is also the retrieval times threshold, and assuming that the detection data of the single target with the deepest retrieval is N k is the number of the k-layer associated target list, and after the retrieval with the retrieval depth k, the obtained detection data set can be denoted as S, and S can be represented by the following formula:

[0139]

[0140] Since Among them, N0 = 1 and X1 = 1. Then, S can be simplified and expressed by the following formula:

[0141]

[0142] Step 2104: Take each obtained detection data as the detection data associated with the feature information to obtain the foregoing detection data.

[0143] In the embodiments of the present application, by performing the foregoing steps 2101 to 2013, each detection data can be retrieved from multiple detection data. Since each retrieved detection data is based on the feature information of the target to be tracked or retrieved based on the target associated with the target that matches the feature information, and is associated with the feature information of the target to be tracked. Therefore, when performing step 2104, each obtained detection data is taken as the detection data associated with the feature information, so as to obtain each detection data associated with the feature information of the target to be tracked.

[0144] Step 220: Determine the spatio-temporal change information of the target to be tracked according to the acquisition time and geographical location included in each detection data.

[0145] In some embodiments of the present application, referring to Figure 5 as shown, when performing step 220, the function of determining the spatio-temporal change information of the target to be tracked can be realized by performing the following steps:

[0146] Step 2201: Add the retrieval cumulative times and feature matching degrees when each detection data is obtained to their respective detection data to obtain each retrieval result data.

[0147] In the embodiments of the present application, after performing step 210 to obtain each detection data associated with the target to be tracked, step 2201 is performed, and the retrieval and feature matching situations are added to each detection data, that is, the retrieval cumulative times k' and the feature matching degree p of the feature information when each detection data is obtained are added to their respective detection data. Any detection data after adding the foregoing information can be expressed as: T'(i, t, s, c, a, T iC , p, k').

[0148] Optionally, in specific implementation, when adding the above information, more dimensional data can also be added according to actual needs to better reflect the information retrieval process, retrieval details, etc.

[0149] In some embodiments of the present application, after performing step 2201 to obtain each detection data after adding the foregoing information, data deduplication can also be performed on each detection data after adding the foregoing information based on the identification information of the target, so as to obtain each retrieval result data of the target to be tracked.

[0150] Optionally, when performing data deduplication, for each detection data with the same identification information, the result with a smaller cumulative retrieval count can be preferentially retained. In this way, among the obtained retrieval result data for the same target, only one detection data is retained, and the detection data with a higher degree of association with the feature information of the target to be tracked is retained, which can improve the accuracy of the subsequent drawn movement trajectory and thus improve the trajectory tracking accuracy rate.

[0151] Step 2202: According to the preset geographical regions, divide the retrieval result data located in the same geographical region among the retrieval result data into the same data set, obtaining multiple data sets.

[0152] In this application, a data aggregation method that aggregates data by geographical location s and time t is used to achieve target trajectory tracking. Among them, the geographical location level can have multiple levels, and the smallest level can be the geographical location of a single acquisition device. The data aggregation level depends on the selected geographical location level, and multiple geographical location levels can also be selected for data aggregation.

[0153] In some embodiments, the tracking area can be pre-divided into different geographical regions. Each geographical region can correspond to a point in the map, such as the geographical location of an acquisition device, or can correspond to a block in the map, such as the area where a street is located, and multiple acquisition devices can be deployed in this area, etc. Then, when performing step 2202, data aggregation is performed on the retrieval result data according to the preset geographical regions, that is, the retrieval result data located in the same geographical region among the retrieval result data are divided into the same data set, thereby obtaining multiple data sets, where each data set includes at least one retrieval result data.

[0154] Step 2203: Based on the acquisition time, geographical location, cumulative retrieval count, and feature matching degree included in the retrieval result data in multiple data sets, respectively determine the confidence levels of the acquisition time and geographical location pairs of the target to be tracked in each data set.

[0155] In some embodiments of this application, when performing step 2203, refer to Figure 6 As shown, the following steps are respectively performed for each data set among the multiple data sets to determine the confidence levels of the acquisition time and geographical location pairs of the target to be tracked in each data set:

[0156] Step 600: Based on the average acquisition time of the retrieval result data in any one of the multiple data sets, and the standard deviation of the acquisition time, determine the splitting index value of this data set.

[0157] In some embodiments of the present application, after obtaining multiple data sets in step 2202, for any data set, when performing step 600, the acquisition times included in each retrieval result data in the data set are obtained, and the average value of the obtained acquisition times is calculated, that is, the average acquisition time of the data set is obtained. Then, the standard deviation of the acquisition time of the data set is calculated, and the ratio of the standard deviation of the acquisition time to the average acquisition time is calculated, and this ratio is the splitting index value of the data set.

[0158] Suppose data set A contains n retrieval result data, and the acquisition time included in any detection result data is denoted as t Ai , where i ∈ [1, n]. Then, the average acquisition time μ of this data set A can be expressed by the following formula:

[0159]

[0160] The standard deviation σ of the acquisition time of this data set A can be expressed by the following formula:

[0161]

[0162] Then, the splitting index value of this data set A =

[0163] Step 610: Determine whether the splitting index value is greater than a preset coefficient of variation constant. If so, execute step 620. If not, execute steps 630 to 640.

[0164] In the embodiments of the present application, after obtaining the splitting index values of each data set in step 600, step 610 is executed to determine whether the splitting index value is greater than a preset coefficient of variation constant, where the coefficient of variation constant can be determined in advance according to experience or adjusted according to actual needs to ensure that the time splitting dimension is within the acceptable range of the application scenario.

[0165] For example, if the application scenario needs to detect the detailed trajectory of a target within a specified area within a short time interval (e.g., within 8 hours), a smaller coefficient of variation constant needs to be set so that the time splitting of this block / point is finer; if the application scenario needs to detect the rough trajectory of a target within a specified area within a long time interval, a larger coefficient of variation constant can be set so that the time splitting of the block / point will be coarser and the trajectory will be coarser.

[0166] In some embodiments, if it is determined that the splitting index value of the data set is greater than the coefficient of variation constant, then step 620 is executed. In other embodiments, if it is determined that the splitting index value of the data set is less than or equal to the coefficient of variation constant, then steps 630 to 640 are executed.

[0167] Step 620: Determine the confidence of the acquisition time and geographical location pair of the target to be tracked in this data set based on the acquisition time, geographical location, cumulative retrieval times, and feature matching degree included in the retrieval result data in this data set.

[0168] In some embodiments, if it is determined that the splitting index value of this data set is greater than the coefficient of variation constant, then perform the subsequent step 620. The confidence P of the acquisition time and geographical location pair of the target to be tracked in this data set can be determined by the following formula, where the acquisition time in the acquisition time and geographical location pair can be the average acquisition time of this data set:

[0169]

[0170] where P is the confidence of the acquisition time and geographical location pair of the target to be tracked in data set A, that is, the confidence that the target to be tracked is located at the geographical location of the retrieval result data i at the average acquisition time μ of this data set A. Data set A is any one of multiple data sets; i is the serial number, i ∈ [1, n]; σ is the standard deviation of the acquisition time of this data set A; n is the number of retrieval result data included in data set A; k is the retrieval times threshold; k Ai is the cumulative retrieval times when the detection data corresponding to the retrieval result data i is obtained; p Ai is the feature matching degree of the retrieval result data i; j is the serial number, j ∈ [1, k].

[0171] It should be noted that in the embodiments of the present application, the higher the confidence, the higher the credibility that the target to be tracked is located at the geographical location of the corresponding detection data at this average acquisition time.

[0172] Step 630: Based on the coefficient of variation constant, split this data set in the time dimension so that the splitting index values corresponding to the split sub-data sets are greater than the coefficient of variation constant.

[0173] In some embodiments, if it is determined that the splitting index value corresponding to this data set is less than or equal to the foregoing coefficient of variation constant, then perform the subsequent step 630. When performing step 630, based on the coefficient of variation constant, perform time splitting on each retrieval result data in this data set to divide the retrieval result data in this data set into sub-data sets in different time intervals to obtain the foregoing respective sub-data sets.

[0174] It should be noted that the present application does not limit the splitting method of the specific data set. That is, any algorithm can be used to split the data set according to the time interval. For example, an optimization algorithm is adopted. Based on the splitting index that the ratio of the standard deviation of the acquisition time of the sub-data sets after splitting to the average acquisition time is greater than the aforementioned coefficient of variation constant, a suitable time interval is first determined. Then, based on the time interval, the retrieval result data in the data set is divided, so as to obtain the aforementioned respective sub-data sets.

[0175] Step 640: Based on the acquisition time, geographical location, retrieval cumulative times, and feature matching degrees included in the retrieval result data in each sub-data set, respectively determine the confidence levels of the acquisition time and geographical location pairs of the target to be tracked in each sub-data set.

[0176] In some embodiments of the present application, when performing step 640, the following operations are respectively performed for each sub-data set in each of the sub-data sets: Based on the acquisition time included in the retrieval result data in any one of the sub-data sets in the respective sub-data sets, the average acquisition time and the standard deviation of the acquisition time corresponding to the sub-data set are obtained. Then, using the aforementioned formula for calculating the confidence level P, based on the average acquisition time, geographical location, retrieval cumulative times of each retrieval result data in the sub-data set, feature matching degree, and retrieval times threshold corresponding to the sub-data set, determine the confidence level of the acquisition time and geographical location pair of the target to be tracked in the sub-data set, where the acquisition time is the average acquisition time of the sub-data set, that is, the confidence level when the target to be tracked is at the average acquisition time of the sub-data set and is located at the geographical location of the corresponding retrieval result data. In this way, the respective confidence levels of the target to be tracked for the data set are obtained.

[0177] Step 2204: Take the acquisition time and geographical location pairs corresponding to the confidence levels greater than or equal to the confidence level threshold among the determined confidence levels as the spatio-temporal change information of the target to be tracked.

[0178] In the embodiments of the present application, a confidence level threshold can be set in advance to screen the obtained respective retrieval result data.

[0179] Optionally, the confidence level threshold can take the value of 1. Then, when performing step 2204, compare the determined confidence levels with the confidence level threshold, select the respective retrieval result data with confidence levels greater than or equal to the confidence level threshold (such as the aforementioned 1), and determine the average acquisition time and geographical location pairs of the selected respective retrieval result data as the spatio-temporal change information of the target to be tracked.

[0180] Step 230: Based on the spatio-temporal change information, draw and display the moving trajectory of the target to be tracked.

[0181] In some embodiments of the present application, refer to Figure 7As shown in the figure, when performing step 230, the following steps can be specifically executed:

[0182] Step 2301: Based on the map data, map the geographical positions in the spatio-temporal change information to the corresponding maps in the map data respectively to obtain each trajectory point;

[0183] Step 2302: Connect each trajectory point in sequence according to each collection time in the spatio-temporal change information to obtain the moving trajectory of the target to be tracked;

[0184] Step 2303: Display the moving trajectory on a preset interface.

[0185] In the embodiment of the present application, based on the time and position information obtained from the retrieval result data, by executing steps 2301 to 2303, the moving trajectory of the target to be tracked is graphically drawn, realizing the macroscopic moving trajectory tracking of the target, and the moving trajectory of the target and the activity range path can be obtained more clearly.

[0186] Next, taking a specific embodiment as an example, a method for tracking the target trajectory in the embodiment of the present application will be illustrated.

[0187] For example, assume that Figure 8 As shown in the schematic diagram of a simple road scene, there are 9 buildings in this road scene, such as Figure 8 shown in the small rectangular areas, and the areas between the small rectangular areas are roads. The road scene is pre-divided into multiple points, such as Figure 9 shown, starting from the upper left corner of the image, they are sequentially denoted as A1, A2, A3, A4,..., A 40 .

[0188] Assume that through the foregoing retrieval method, based on the feature information of the target to be tracked T, each detection data retrieved from multiple detection data according to a retrieval depth of 2 is respectively added with the retrieval cumulative number and the feature matching degree obtained when it is obtained, and data deduplication is performed based on the identification information of the target to obtain each retrieval result data; then, according to the preset geographical blocks, the retrieval result data located in the same geographical block are divided into the same data set, that is, data aggregation is performed to obtain multiple data sets; furthermore, further time aggregation can be performed to obtain each data set and sub-data set to be processed.

[0189] Also assume that after the foregoing processing, the geographical positions in the obtained data set and / or sub-data set are sequentially A 13 , A 14 , A 20 , A 25 , A 26 , A 27 , A 28 , A 29 , A33 , A7, A9, A 23 ; Then, using the foregoing method for calculating the confidence level, that is, the formula for P above, the confidence levels of each point are calculated as shown in the following table:

[0190] Table of confidence levels and average acquisition times for each point

[0191] Point location (geographical location) Confidence level (P) Average acquisition time (μ) sorting <![CDATA[A7]]> 0.37 7 <![CDATA[A9]]> 0.86 4 <![CDATA[A 13 > 2.11 1 <![CDATA[A 14 > 3.21 2 <![CDATA[A 20 > 1.21 3 <![CDATA[A 23 > 0.56 5 <![CDATA[A 25 > 0.92 6 <![CDATA[A 26 > 1.52 8 <![CDATA[A 27 > 1.86 9 <![CDATA[A 28 > 5.33 10 <![CDATA[A 29 > 2.45 11 <![CDATA[A 33 > 2.11 12

[0192] Assume that the points with P≥1 are still selected, then the selected points are successively (A 13 , 1), (A 14 , 2), (A 20 , 3), (A 26 , 8), (A 27 , 9), (A 28 , 10), (A 29 , 11), (A 33 , 12); Then, execute step 2301 to step 2303, map the geographical locations of each point to the corresponding map in the map data, and connect them in sequence according to the average acquisition time to obtain the movement trajectory of the target A to be tracked as shown in Figure 10 the movement trajectory of the target A to be tracked shown in

[0193] In some embodiments of the present application, after executing step 210 and before executing step 220, the following operations 1 to 3 can also be performed to enrich the retrieved detection data: Operation 1, in response to a display operation for candidate detection data associated with any trajectory point in the movement trajectory, display the candidate detection data and the associated candidate base image; Operation 2, if an additional retrieval instruction for the associated target in the candidate base image is received, then based on the additional retrieval instruction, retrieve the detection data of the associated target from multiple detection data, where the additional retrieval instruction includes the associated target, which is a target strongly related to the target to be tracked included in the image data; Operation 3, add the retrieved detection data to the foregoing detection data, and redraw the movement trajectory of the target to be tracked based on the newly obtained detection data.

[0194] Among them, the foregoing associated target is included in the candidate base image and is a target strongly related to the target to be tracked included in the image data. For example, in the foregoing example, a person rides a bicycle. If the target to be tracked is a person and the bicycle also exists in the candidate base image, when it is manually confirmed that the bicycle in the candidate base image is the bicycle ridden by the target to be tracked in the image data, the operator can trigger an additional retrieval instruction for the associated target of the bicycle.

[0195] Still as shown in Figure 10 , the target A to be tracked is at A 20 , A 26There is a track loss. The user can click on Figure 10 any track point in the shown movement track, trigger an indication to display the retrieval result data of this track point, and then, based on the aforementioned operation 1, display the candidate detection data associated with this track point and the associated candidate basic image. If the user determines that although the target to be tracked does not exist in the candidate basic image, there is a target strongly related to the target to be tracked. For example, in the aforementioned example, the target to be tracked is a person, and the person is riding a bicycle. Then, the bicycle is associated with the target to be tracked. Exactly in the candidate basic image, although the person who is the target to be tracked does not exist, there is a bicycle strongly related to it. Then, it is determined that an additional retrieval instruction needs to be triggered.

[0196] After receiving the additional retrieval instruction, the central data node performs operation 2. Based on the additional retrieval instruction, in multiple detection data, it retrieves the associated target included in the additional retrieval instruction. At this time, the retrieval cumulative count of the candidate detection data is adjusted to 1. Specifically, when performing operation 2, based on the feature information of the associated target, from multiple detection data, it obtains the detection data of the reference target that matches the feature information of the associated target; then, from multiple detection data, it obtains the detection data corresponding to the identification information of each target included in the detection data of the reference target. Finally, each obtained detection data is respectively determined as the detection data associated with the associated target.

[0197] Continuing with the aforementioned example, when performing operation 2, based on the feature information of the bicycle, it retrieves for the bicycle from multiple detection data, thereby obtaining the detection data that matches the feature information of the bicycle, and the detection data corresponding to the identification information of each target included in the detection data. Then, each obtained detection data is respectively determined as the detection data associated with the bicycle. Then, when performing operation 3, the retrieved detection data associated with the bicycle is added to the aforementioned each detection data, and based on the newly obtained each detection data, the movement track is redrawn. The redrawn track can be as Figure 11 shown.

[0198] Refer to Figure 10 and Figure 11 , it can be found by comparison that by increasing the retrieval target, that is, for the associated target in the candidate basic image of the aforementioned track point, further retrieving in multiple detection data, the samples for drawing the movement track of the target to be tracked can be enriched, and the tracking accuracy can be improved.

[0199] In some embodiments, the image data containing the target to be tracked obtained by performing step 200 can be collected by the registered acquisition device registered in the target track tracking system in the embodiments of the present application, or can be collected by an unregistered acquisition device. Specifically, it can be identified through the identification information of the acquisition device carried by the image data.

[0200] If the identification information of the acquisition device carried by the foregoing image data is the identification information of a registered acquisition device, the device geographical location corresponding to the identification information of the acquisition device carried by the image data can also be obtained; connect the last trajectory point in the movement trajectory with the trajectory point of the device geographical location on the map to obtain and display the final movement trajectory of the target to be tracked.

[0201] After introducing how to draw the movement trajectory of the target to be tracked using multiple detection data, it should also be noted that since the foregoing data collection process can also collect real-time and continuously updated data, then, assuming that the image data obtained in step 200 is real-time and continuously updated data, then, after obtaining the feature information of the target to be tracked in step 200, the foregoing target detection and feature association method can also be used to obtain the association relationship between the targets in the image data, construct at least one detection data based on the image data, and store it, and this data can also be used for subsequent retrieval and drawing of the movement trajectory.

[0202] Based on the same inventive concept, refer to Figure 12 As shown, an object trajectory tracking device is provided in an embodiment of the present application, including:

[0203] An object detection module 1210, configured to obtain image data including the object to be tracked, perform object detection on the image data, and obtain the feature information of the object to be tracked;

[0204] A retrieval module 1220, configured to perform at least one retrieval in a plurality of detection data based on the feature information to obtain each detection data associated with the feature information, where any one of the plurality of detection data includes the identification information of the target, the feature information of the target in the base image associated with the any one of the detection data, the identification information of the associated targets, and the acquisition time of the base image and the geographical location of the acquisition device;

[0205] A data aggregation module 1230, configured to determine the spatio-temporal change information of the object to be tracked according to the acquisition time and geographical location included in each detection data;

[0206] A display module 1240, configured to draw and display the movement trajectory of the object to be tracked based on the spatio-temporal change information.

[0207] In a possible implementation manner, the retrieval module 1220 is specifically configured to:

[0208] Based on the feature information, obtain the detection data of the target that matches the feature information from the plurality of detection data, and the detection data corresponding to the identification information of each target included in the detection data;

[0209] When it is determined that the retrieval times threshold has not been reached, repeat the following operations until the retrieval times threshold is reached:

[0210] For any retrieval, each currently obtained target is used as a retrieval object. For any retrieval object, based on the feature information of the any retrieval object, from the multiple detection data, the detection data of candidate targets that match the feature information of the any retrieval object is obtained, and

[0211] the detection data corresponding to the identification information of each target included in the detection data of the candidate target;

[0212] Each obtained detection data is used as the detection data associated with the feature information to obtain the respective detection data.

[0213] In a possible implementation manner, the data aggregation module 1230 is specifically configured to:

[0214] The retrieval cumulative times and feature matching degrees when each detection data is obtained are respectively added to the respective detection data to obtain respective retrieval result data;

[0215] According to the preset geographical regions, the retrieval result data located in the same geographical region in the respective retrieval result data are divided into the same data set to obtain multiple data sets;

[0216] Based on the collection time, geographical location, retrieval cumulative times, and feature matching degrees included in the retrieval result data in the multiple data sets, the confidence degrees of the collection time and geographical location pairs of the target to be tracked in each data set are respectively determined;

[0217] The collection time and geographical location pairs corresponding to the confidence degrees greater than or equal to the confidence threshold among the determined confidence degrees are used as the spatio-temporal change information of the target to be tracked.

[0218] In a possible implementation manner, by performing the following operations, the confidence degrees of the collection time and geographical location pairs of the target to be tracked in each data set are determined:

[0219] Based on the average collection time of the retrieval result data in any data set and the standard deviation of the collection time, the splitting index value of the any data set is determined;

[0220] If it is determined that the splitting index value is greater than the preset coefficient of variation constant, then based on the collection time, geographical location, retrieval cumulative times, and feature matching degrees included in the retrieval result data in the any data set, the confidence degrees of the collection time and geographical location pairs of the target to be tracked in the any data set are determined;

[0221] If it is determined that the split index value is less than or equal to the coefficient of variation constant, then based on the coefficient of variation constant, the any dataset is split in the time dimension such that the split index values corresponding to the split sub-datasets are greater than the coefficient of variation constant; based on the collection time, geographical location, retrieval cumulative count, and feature matching degree included in the retrieval result data within each sub-dataset, the confidence degrees of the collection time and geographical location pair of the target to be tracked in each sub-dataset are respectively determined.

[0222] In a possible implementation manner, the display module 1240 is specifically configured to:

[0223] Based on the map data, map each geographical location in the spatio-temporal change information to the corresponding map in the map data respectively to obtain each trajectory point;

[0224] Connect the respective trajectory points in sequence according to each collection time in the spatio-temporal change information to obtain the movement trajectory of the target to be tracked;

[0225] Display the movement trajectory in a preset interface.

[0226] In a possible implementation manner, after displaying the movement trajectory, the display module 1240 is further configured to:

[0227] In response to a display operation for candidate detection data associated with any trajectory point in the movement trajectory, display the candidate detection data and the associated candidate base image;

[0228] If an additional retrieval instruction for an associated target in the candidate base image is received, then based on the additional retrieval instruction, retrieve the detection data associated with the associated target from the multiple detection data, where the additional retrieval instruction includes the associated target, which is a target strongly correlated with the target to be tracked included in the image data;

[0229] Add the retrieved detection data to the respective detection data, and redraw the movement trajectory of the target to be tracked based on the newly obtained respective detection data.

[0230] In a possible implementation manner, the display module 1240 is specifically configured to:

[0231] Based on the feature information of the associated target, obtain the detection data of the reference target that matches the feature information of the associated target from the multiple detection data;

[0232] Obtain the detection data corresponding to the identification information of each target included in the detection data of the reference target from the multiple detection data;

[0233] Each of the obtained detection data is respectively determined as the detection data associated with the associated target.

[0234] In a possible implementation manner, if the identification information of the acquisition device carried in the image data is the identification information of a registered acquisition device, the display module 1240 is further configured to:

[0235] Obtain the device geographical location corresponding to the identification information of the acquisition device carried in the image data;

[0236] Connect the last track point in the movement track with the track point of the device geographical location on the map, and obtain and display the final movement track of the target to be tracked.

[0237] Refer to Figure 13 As shown, in an embodiment of the present application, a server is provided, and this server can implement the functions of the foregoing target track tracking method. Refer to Figure 13 and this server includes:

[0238] At least one processor 131, and a memory 132 connected to at least one processor 131. In the embodiment of the present application, the specific connection medium between the processor 131 and the memory 132 is not limited. Figure 13 In is taken as an example that the processor 131 and the memory 132 are connected through a bus 130. The bus 130 is Figure 13 shown in thick lines in, and the connection manners between other components are only for illustrative purposes and are not to be taken as a limitation. The bus 130 can be divided into an address bus, a data bus, a control bus, etc. For the convenience of representation, Figure 13 only one thick line is shown in, but it does not mean that there is only one bus or one type of bus. Alternatively, the processor 131 can also be referred to as a controller, and the name is not limited.

[0239] In the embodiment of the present application, the memory 132 stores instructions executable by at least one processor 131. By executing the instructions stored in the memory 132, at least one processor 131 can execute the target track tracking method described above. The processor 131 can implement the functions of the target track tracking device.

[0240] In a possible design, the processor 131 may include one or more processing units. The processor 131 may integrate an application processor and a modulation and demodulation processor. Among them, the application processor mainly processes the operating system, user interface, application programs, etc., and the modulation and demodulation processor mainly processes wireless communication. It can be understood that the above modulation and demodulation processor may not be integrated into the processor 131. In some embodiments, the processor 131 and the memory 132 can be implemented on the same chip, and in some embodiments, they can also be separately implemented on independent chips.

[0241] The processor 131 may be a general-purpose processor. For example, it can be a Central Processing Unit (CPU), a digital signal processor, an application-specific integrated circuit, a field-programmable gate array, or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, which can implement or execute the various methods, steps, and logic block diagrams disclosed in the embodiments of the present application. The general-purpose processor can be a microprocessor or any conventional processor, etc. The steps of the target trajectory tracking method disclosed in combination with the embodiments of the present application can be directly implemented by the execution of the hardware processor, or can be implemented by the combination of the hardware and software modules in the processor.

[0242] The memory 132, as a non-volatile computer-readable storage medium, can be used to store non-volatile software programs, non-volatile computer-executable programs, and modules. The memory 132 can include at least one type of storage medium. For example, it can include flash memory, hard disk, multimedia card, card-type memory, Random Access Memory (RAM), Static Random Access Memory (SRAM), Programmable Read Only Memory (PROM), Read Only Memory (ROM), Electrically Erasable Programmable Read-Only Memory (EEPROM), magnetic memory, magnetic disk, optical disk, etc. The memory 132 is any other medium that can be used to carry or store the desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited thereto. The memory 132 in the embodiments of the present application can also be a circuit or any other device capable of implementing a storage function, for storing program instructions and / or data.

[0243] By designing and programming the processor 131, the code corresponding to the target trajectory tracking method introduced in the foregoing embodiments can be solidified into the chip, so that the chip can execute the steps of the target trajectory tracking method in the foregoing embodiments when running. How to design and program the processor 131 is a well-known technology to those skilled in the art and will not be elaborated here.

[0244] Based on the same inventive concept, the embodiments of the present application provide a computer-readable storage medium, when the instructions in the storage medium are executed by a processor, enabling the processor to execute any one of the methods in the above various embodiments.

[0245] In some possible embodiments, various aspects of the target trajectory tracking method provided in this application can also be implemented in the form of a program product, which includes program code. When the program product runs on a device, the program code is used to cause the device to execute the steps in the methods according to various exemplary embodiments described above in this specification.

[0246] Those skilled in the art should understand that the embodiments of this application can be provided as a method, a system, or a computer program product. Therefore, this application can take the form of an all-hardware embodiment, an all-software embodiment, or an embodiment combining software and hardware aspects. Moreover, this application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memory, CD-ROM, optical memory, etc.) containing computer-usable program code.

[0247] This application is described with reference to the flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to this application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, as well as the combination of flows and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate a device for implementing the functions specified in one or more flows in the flowchart and / or one or more blocks in the block diagram.

[0248] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory generate a manufactured product including an instruction device, and the instruction device implements the functions specified in one or more flows in the flowchart and / or one or more blocks in the block diagram.

[0249] These computer program instructions can also be loaded onto a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process. Thus, the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one or more flows in the flowchart and / or one or more blocks in the block diagram.

[0250] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Thus, if these modifications and variations of this application fall within the scope of the claims of this application and their equivalent technologies, this application is also intended to include these modifications and variations.

Claims

1. A target trajectory tracking method, characterized in that, Including: Obtain image data including the target to be tracked, perform target detection on the image data, and obtain the feature information of the target to be tracked; Based on the feature information, perform at least one retrieval among multiple detection data to obtain each detection data associated with the feature information, where any one of the multiple detection data includes the identification information of the target, the feature information of the target in the base image associated with the any one of the detection data, and the identification information of each associated target, as well as the acquisition time of the base image and the geographical location of the acquisition device; According to the acquisition time and geographical location included in each detection data, determine the spatio-temporal change information of the target to be tracked; Based on the spatio-temporal change information, draw the movement trajectory of the target to be tracked and display it.

2. The method according to claim 1, wherein The performing at least one retrieval among multiple detection data based on the feature information to obtain each detection data associated with the feature information includes: Based on the feature information, obtain the detection data of the target that matches the feature information from the multiple detection data, and the detection data corresponding to the identification information of each target included in the detection data; When it is determined that the retrieval times threshold is not reached, repeat the following operations until the retrieval times threshold is reached: For any one retrieval, use each target obtained currently as the retrieval object. For any one retrieval object, based on the feature information of the any one retrieval object, obtain the detection data of the candidate target that matches the feature information of the any one retrieval object from the multiple detection data, and The detection data corresponding to the identification information of each target included in the detection data of the candidate target; Use each obtained detection data as the detection data associated with the feature information to obtain the each detection data.

3. The method according to claim 1 or 2, characterized in that, The determining the spatio-temporal change information of the target to be tracked according to the acquisition time and geographical location included in each detection data includes: Add the retrieval cumulative times and feature matching degrees when each detection data is obtained to the respective detection data to obtain each retrieval result data; According to the preset geographical blocks, divide the retrieval result data located in the same geographical block among the each retrieval result data into the same data set to obtain multiple data sets; Based on the acquisition time, geographical location, retrieval cumulative times, and feature matching degrees included in the retrieval result data in the multiple data sets, respectively determine the confidence degrees of the acquisition time and geographical location pair of the target to be tracked in each data set; Use the acquisition time and geographical location pair corresponding to the confidence degrees greater than or equal to the confidence threshold among the determined each confidence degree as the spatio-temporal change information of the target to be tracked.

4. The method according to claim 3, characterized in that By performing the following operations, determine the confidence degrees of the acquisition time and geographical location pair of the target to be tracked in each data set: Based on the average acquisition time of the retrieval result data in any one data set, and the acquisition time standard deviation, determine the splitting index value of the any one data set; If it is determined that the split index value is greater than a preset coefficient of variation constant, then based on the collection time, geographical location, cumulative search times, and feature matching degree included in the retrieval result data in any one of the data sets, determine the confidence level of the collection time and geographical location pair of the target to be tracked in any one of the data sets; If it is determined that the split index value is less than or equal to the coefficient of variation constant, then based on the coefficient of variation constant, split any one of the data sets in the time dimension so that the split index values corresponding to the split sub-data sets are greater than the coefficient of variation constant; based on the collection time, geographical location, cumulative search times, and feature matching degree included in the retrieval result data in each sub-data set, determine the confidence level of the collection time and geographical location pair of the target to be tracked in each sub-data set respectively.

5. The method according to claim 1 or 2, characterized in that, The drawing and display of the movement track of the target to be tracked based on the spatio-temporal change information include: Based on the map data, map each geographical location in the spatio-temporal change information to the corresponding map in the map data respectively to obtain each track point; Connect each track point in sequence according to each collection time in the spatio-temporal change information to obtain the movement track of the target to be tracked; Display the movement track in a preset interface.

6. The method according to claim 1 or 2, characterized in that, After displaying the movement track, the method further includes: In response to a display operation on the candidate detection data associated with any track point in the movement track, display the candidate detection data and the associated candidate base image; If an additional retrieval instruction for the associated target in the candidate base image is received, then based on the additional retrieval instruction, retrieve the detection data associated with the associated target from the multiple detection data, where the additional retrieval instruction includes the associated target, which is a target strongly correlated with the target to be tracked included in the image data; Add the retrieved detection data to the respective detection data, and redraw the movement track of the target to be tracked based on the newly obtained respective detection data.

7. The method according to claim 6, wherein The retrieving of the detection data associated with the associated target from the multiple detection data based on the additional retrieval instruction includes: Based on the feature information of the associated target, obtain the detection data of the reference target that matches the feature information of the associated target from the multiple detection data; From the multiple detection data, obtain the detection data corresponding to the identification information of each target included in the detection data of the reference target; Determine each obtained detection data as the detection data associated with the associated target respectively.

8. The method according to claim 1 or 2, characterized in that, If the identification information of the acquisition device carried by the image data is the identification information of a registered acquisition device, then the method further includes: Obtain the device geographical location corresponding to the identification information of the acquisition device carried by the image data; Connect the last track point in the movement track with the track point of the device geographical location in the map to obtain and display the final movement track of the target to be tracked.

9. A target trajectory tracking device, characterized in that, including: A target detection module, configured to obtain image data containing a target to be tracked, perform target detection on the image data, and obtain feature information of the target to be tracked; A retrieval module, configured to perform at least one retrieval among a plurality of detection data based on the feature information, and obtain each detection data associated with the feature information, wherein any one of the plurality of detection data includes identification information of a target, feature information of the target in a base image associated with the any one of the detection data, identification information of each associated target, and the acquisition time of the base image and the geographical location of the acquisition device; A data aggregation module, configured to determine spatio-temporal change information of the target to be tracked according to the acquisition time and geographical location included in each detection data; A display module, configured to draw and display a movement trajectory of the target to be tracked based on the spatio-temporal change information.

10. A server, characterized in that, Comprising: A memory, configured to store computer programs or instructions; A processor, configured to execute the computer programs or instructions in the memory, so that the method according to any one of claims 1-8 is executed.

11. A computer-readable storage medium, characterized in that, When the instructions in the storage medium are executed by the processor, the processor is enabled to execute the method according to any one of claims 1-8.

Images