Method and apparatus for finding a target

Abstract

The application discloses a method and device for searching for a target. The method for searching for the target comprises the following steps: determining a plurality of sampling points in historical trajectory data of a wearable device of the target; searching for all video acquisition devices within a preset range of each sampling point; pushing video data of each video acquisition device based on the intersection degree of the monitoring range of each searched video acquisition device and the historical trajectory data, so as to search for the target based on the pushed video data, and the pushing sequence of the video data of the video acquisition device is positively correlated with the corresponding intersection degree of the video acquisition device. The application can reduce video checking time and improve the searching efficiency for the target.

Description

Technical Field

[0001] This application relates to the field of video processing technology, and in particular to a method and apparatus for finding targets. Background Technology

[0002] With the increasing development of transportation and the rise of the floating population, the number of lost people and pets is on the rise. Lost targets are often not quickly recovered due to inefficient search methods, or even disappear without a trace. Therefore, how to utilize new technologies for rapid and effective target retrieval urgently needs to be addressed. Cities are equipped with a large number of video surveillance devices; in first-tier cities like Guangzhou and Shanghai, the number of cameras has reached hundreds of thousands, covering all important urban areas. These cameras capture a large amount of information about pedestrians and objects in real time. How to effectively utilize this video surveillance to automate the search for people and objects in cities will become the next hot topic.

[0003] Currently, the primary methods for retrieving targets are manual inspection or the use of artificial intelligence algorithms to identify target features from massive amounts of video footage. However, the investigation and processing of massive amounts of real-world video footage often requires enormous effort, manpower, and resources to pinpoint the target's location. When clues are missing or existing data is not fully utilized, the time and geographical scope of the incident can only be roughly estimated through inference, causing the number of real-world videos requiring investigation to increase rapidly in both time and location dimensions. Furthermore, as the investigation period lengthens, the search area also expands, and the workload of reviewing videos increases dramatically. Summary of the Invention

[0004] This application provides a method and apparatus for finding targets, which can reduce video screening time and improve the efficiency of target finding.

[0005] To achieve the above objectives, this application provides a method for finding a target, the method comprising:

[0006] Identify several sampling points in the historical trajectory data of the wearable device used to determine the target;

[0007] Search all video capture devices within a preset range at each sampling point;

[0008] Based on the monitoring range and historical trajectory data of each video acquisition device found, video data from each video acquisition device is pushed to facilitate target location based on the pushed video data. The order in which video data is pushed from each video acquisition device is positively correlated with the degree of intersection between the corresponding video acquisition devices.

[0009] This includes pushing video data from each video acquisition device based on the overlap between the monitoring range and historical trajectory data of the various video acquisition devices found in the search, including:

[0010] Based on the monitoring range and historical trajectory data of each video acquisition device, the importance of each video acquisition device is determined. The importance of the video acquisition device is positively correlated with the degree of intersection of the video acquisition device.

[0011] Video data from each video acquisition device is pushed out in descending order of importance.

[0012] The importance of each video acquisition device is determined based on the overlap between its monitoring range and historical trajectory data. This includes:

[0013] Based on the degree of intersection of each video acquisition device and the acquisition time of the sampling points of each video acquisition device, the importance of each video acquisition device is determined. The importance of a video acquisition device is negatively correlated with the acquisition time of the sampling points of the corresponding video acquisition device.

[0014] The importance of each video acquisition device is determined based on the overlap between its monitoring range and historical trajectory data. This includes:

[0015] The monitoring range of each video acquisition device is determined based on the parameters of each video acquisition device.

[0016] The length of the wearable device's trajectory line within the monitoring range of each video acquisition device is determined, and the trajectory line is determined based on historical trajectory data;

[0017] Calculate the ratio of the length corresponding to each video acquisition device to the total length of the trajectory line, and use the ratio as the degree of intersection of each video acquisition device.

[0018] The parameters of the video acquisition devices include their coordinates, orientation angle, viewing angle, and depth of field. The monitoring range of each video acquisition device is determined based on its parameters, including:

[0019] Using the coordinates of each video acquisition device as the center, the depth of field of each video acquisition device as the radius, and the orientation angle and viewing angle of each video acquisition device as the range, the fan-shaped monitoring range of each video acquisition device is determined.

[0020] This includes searching all video capture devices within a preset range at each sampling point, previously including:

[0021] Collect and store parameters of all video capture devices;

[0022] If the depth of field of the video acquisition device cannot be calculated during the process of collecting parameters of the video acquisition device, the theoretical depth of field value determined based on the model of the video acquisition device will be used as the depth of field of the video acquisition device.

[0023] This includes pushing video data from various video capture devices, followed by:

[0024] If the pushed video data cannot meet the search requirements, adjust the preset range and / or the interval between two adjacent sampling points;

[0025] Based on the adjusted preset range and / or the time interval between two adjacent sampling points, return several sampling points from the historical trajectory data of the wearable device that performed the target determination.

[0026] Among them, several sampling points in the historical trajectory data of the wearable device used to determine the target include:

[0027] Historical trajectory data is sampled at preset time intervals to obtain several sampling points in the historical trajectory data.

[0028] To achieve the above objectives, this application also provides an encoder that includes a processor; the processor is configured to execute instructions to implement the steps of the above method.

[0029] To achieve the above objectives, this application also provides a computer-readable storage medium for storing instruction / program data that can be executed to implement the above methods.

[0030] This application first acquires historical trajectory data collected by the wearable device worn by the target, and determines several sampling points in the historical trajectory data of the target's wearable device. Then, it searches for all video acquisition devices within a preset range of each sampling point. Subsequently, based on the monitoring range of the searched video acquisition devices and the degree of intersection with the historical trajectory data, it pushes video data from each video acquisition device to find the target based on the pushed video data. The degree of intersection between the video acquisition devices is positively correlated with the order in which the video acquisition devices are pushed. In this way, the video data collected by the video acquisition devices with a high degree of intersection is pushed first, so that the video data most likely to appear of the target can be pushed first, which can speed up the search for the target based on the pushed video data, reduce video screening time, and thus improve the efficiency of finding the target. Attached Figure Description

[0031] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0032] Figure 1This is a flowchart illustrating one implementation method of the target-finding method of this application;

[0033] Figure 2 This is a schematic diagram illustrating the method for finding a target in this application, specifically the search for video capture devices.

[0034] Figure 3 This is a schematic diagram illustrating the method for determining the monitoring range of the video acquisition device in the target locating process described in this application;

[0035] Figure 4 This is a schematic diagram illustrating the method for determining the degree of intersection of video acquisition devices in the target-finding process described in this application.

[0036] Figure 5 This is a flowchart illustrating another implementation of the method for finding targets in this application;

[0037] Figure 6 This is a schematic diagram of the parameters of the video acquisition device in the method for finding the target in this application;

[0038] Figure 7 This is a schematic diagram of one embodiment of the target-finding device of this application;

[0039] Figure 8 This is a schematic diagram of one embodiment of the computer-readable storage medium of this application. Detailed Implementation

[0040] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application. In addition, unless otherwise specified (e.g., "or additionally" or "or in alternatives"), the term "or" as used herein refers to a non-exclusive "or" (i.e., "and / or"). Furthermore, the various embodiments described herein are not necessarily mutually exclusive, as some embodiments can be combined with one or more other embodiments to form new embodiments.

[0041] With the widespread adoption of smart wearable devices, for purposes such as recording movement trajectories and monitoring the wearer, these devices can provide high-precision global location recording capabilities. By integrating historical location data at each moment, the device's global location historical trajectory information can be obtained. This global location historical trajectory information can be a set of elements consisting of multiple timestamps within a recording period, the latitude and longitude coordinates of the device's global location at the time represented by each timestamp.

[0042] In conventional target location and retrieval solutions, global latitude and longitude information provided by wearable devices can be used to confirm the location indicated by the coordinates provided by the wearable device, thus locating the target. However, if the wearable device and the target become separated, and the goal is to quickly trace their trajectory and whereabouts, it is necessary to search all real-world video footage within the area where the event occurred. This requires manual inspection or the use of artificial intelligence algorithms to identify target features from a large volume of video footage to achieve the retrieval objective. However, the review and processing of massive amounts of real-world video footage often requires enormous effort, manpower, and resources to pinpoint the wearable target's location. When clues are missing or existing data is not fully utilized, only inference can be used to roughly define the time and geographical scope of the unexpected event, causing the number of real-world videos requiring review to increase rapidly in both time and location dimensions. Furthermore, as the time period requiring review increases, the search area also expands, and the workload of reviewing videos increases dramatically.

[0043] Based on this, this application proposes a method for finding targets, which integrates historical trajectory data recorded by wearable devices with information from video acquisition devices, greatly reducing the workload of video retrieval and improving retrieval efficiency.

[0044] Specifically, such as Figure 1 As shown, the method for finding a target proposed in this application specifically includes the following steps. It should be noted that the step numbers are for simplification only and are not intended to limit the execution order of the steps. The execution order of each step in this embodiment can be arbitrarily changed without departing from the technical concept of this application.

[0045] S101: Determine several sampling points in the historical trajectory data of the wearable device of the target.

[0046] First, historical trajectory data collected by the wearable device worn by the target can be obtained, and several sampling points in the historical trajectory data of the target's wearable device can be determined. This allows for subsequent searching of all video capture devices within a preset range of each sampling point. Then, based on the monitoring range of the searched video capture devices and the degree of intersection with the historical trajectory data, video data from each video capture device is pushed to facilitate subsequent target location based on the pushed video data. The degree of intersection between the video capture devices is positively correlated with the order in which the video capture devices are pushed. This prioritizes pushing video data from video capture devices with high intersection, ensuring that the video data most likely to appear for the target is pushed first, thus speeding up the target location process.

[0047] Wearable devices can have the functions of recording location information and data sharing. In certain scenarios that require historical trajectory data of wearable devices, such as when the target has encountered an accident, historical trajectory data can be retrieved by quickly accessing the historical trajectory data from the shared device provided by the target, or by obtaining historical trajectory data from the available wearable devices.

[0048] Historical trajectory data refers to the set of locations collected by a wearable device at all collection moments within a recording period. Thus, the historical trajectory data of a wearable device can consist of multiple original collection points, each corresponding one-to-one with a collection moment, and the location information of the original collection point corresponding to each collection moment is the location collected by the wearable device at each collection moment.

[0049] In one feasible approach, each raw acquisition point of the historical trajectory data from the wearable device can be directly used as each sampling point of the historical trajectory data.

[0050] In another feasible approach, considering the large number of original data points in the acquired historical trajectory data (potentially one travel position per time interval, which can be very short), the resulting trajectory data is extremely large, making subsequent calculations very complex. Therefore, the historical trajectory data can be sampled, and several sampling points can be obtained based on the original data points. This reduces the number of sampling points, lowers data complexity, and reduces the computational complexity of the proposed method.

[0051] Specifically, historical trajectory data can be sampled at preset time intervals to obtain several sampling points from the historical trajectory data. This results in a set of sampling location information points consisting of multiple sampling timestamps and their location information elements. Specifically, the preset time interval can be pre-set; it can be 1 minute, 1 hour, or other time units, without limitation here. That is, in this feasible method, several sampling points can be extracted sequentially from the time period between the start timetamp and the end timetamp in the historical trajectory data according to the set collection step size (i.e., the aforementioned preset time interval).

[0052] S102: Search for all video acquisition devices within a preset range at each sampling point.

[0053] After determining several sampling points in the historical trajectory data of the target wearable device based on step S101, all video acquisition devices within a preset range of each sampling point can be searched.

[0054] Optionally, before step S102, the location of each video acquisition device can be recorded and stored. Thus, in step S102, based on the location of each video acquisition device and the location information of the sampling points, it can be confirmed whether each video acquisition device is within the preset range of the sampling points, thereby identifying all video acquisition devices within the preset range of each sampling point.

[0055] For example, such as Figure 2 As shown, the preset range can be a circle with the position of each sampling point as the center and the preset radius as the radius. In this way, the distance between each video acquisition device and each sampling point can be calculated based on the location of each video acquisition device and the position information of each sampling point. Then, it can be confirmed whether the distance between each video acquisition device and each sampling point is less than the preset radius. All video acquisition devices whose distance from each sampling point is less than the preset radius are regarded as all video acquisition devices within the preset range of each sampling point.

[0056] In a specific example, several sampling points in the historical trajectory data can be traversed sequentially to determine all video acquisition devices within a preset range of each sampling point. When traversing the current sampling point, the global coordinates (latitude and longitude) of that point are obtained, and a circle is drawn with that point as the center (i.e.,...). Figure 2 Point O in the image), with the search range of the video capture device as the radius (i.e., point O). Figure 2 A circular search area plane is drawn using x(m) in the image. From the video acquisition device database, the global coordinate latitude and longitude data of the video acquisition devices are retrieved, and all video acquisition devices within the search area plane are recorded. The key parameters of each video acquisition device are then retrieved from the data. This allows us to obtain the location and device parameters of several video acquisition devices defined by the sampling position and the set search area at a given time. For example... Figure 2 As shown, the four video acquisition devices marked in black fall within the search range plane of the sampling point. In this case, the relevant parameters of these four video acquisition devices can be recorded to determine the monitoring range of the video acquisition devices and the degree of intersection of historical trajectory data.

[0057] S103: Based on the monitoring range and historical trajectory data of each video acquisition device found, push video data from each video acquisition device to find the target based on the pushed video data.

[0058] After searching for all video acquisition devices within the preset range of each sampling point, the monitoring range and historical trajectory data of each video acquisition device can be determined. Then, based on the intersection of the monitoring range and historical trajectory data of each video acquisition device, video data from each device can be pushed to the system. The intersection degree of the video acquisition devices is positively correlated with the order in which the data is pushed. Thus, when searching for a target in the video data according to the pushing order, it is possible to first determine whether there is a target in the video data with a high degree of intersection. Since the probability of a target being present in video data with a high degree of intersection is relatively high, this solution allows for the initial determination of whether a target is present in video data with a high probability of existence, thereby accelerating the search for the target.

[0059] In one feasible approach, the degree of intersection between the various video capture devices can refer to the length of the wearable device's trajectory line falling within the monitoring range of each video capture device. The trajectory line can be determined based on the wearable device's historical trajectory data.

[0060] In this case, the trajectory line can be determined based on the historical trajectory data of the wearable device; and the monitoring range of each video acquisition device can be determined based on the parameters of each video acquisition device; then the length of the wearable device's trajectory line falling within the monitoring range of each video acquisition device can be determined.

[0061] Optionally, the parameters of the video acquisition devices may include data such as the coordinates, orientation angle, viewing angle, and depth of field of the video acquisition devices. This allows for the determination of a sector-shaped monitoring range for each video acquisition device, using its coordinates as the center, its depth of field as the radius, and its orientation angle and viewing angle. For example, such as... Figure 3 As shown, the parameters of each video acquisition device are iterated sequentially, and the viewing angle coverage of the video acquisition device is modeled as a sector. Specifically, the global positioning latitude and longitude coordinates of the video acquisition device are used as the center, the depth of field is the radius of the sector, and the viewing angle of the video acquisition device is the size of the apex of the sector. The orientation of the sector is drawn on the search range plane according to the angle between the orientation data obtained from the orientation data of the video acquisition device database and due north. In this way, the monitoring range of each video acquisition device is determined.

[0062] One method for determining the trajectory line based on the historical trajectory data of wearable devices is to connect all the original collection points in the historical trajectory data in sequence to obtain the trajectory line.

[0063] Of course, in other embodiments, for each sampling point, all original sampling points in the historical trajectory data within the time period from "sampling point acquisition time - first time" to "sampling point acquisition time + second time" can be sequentially connected to obtain a trajectory line. That is, each sampling point can correspond to a trajectory line. In this embodiment, the position information of the sampling point within the time before and after its movement in the historical trajectory data of the wearable device is obtained, and each position point is connected to form a broken line, forming... Figure 4 The image shows the trajectory of the wearable device within a specified time range. Optionally, the sum of the first time and the second time can be less than or equal to a preset unit time. Of course, the sum of the first time and the second time can also be greater than the preset unit time.

[0064] The specific length of the wearable device's trajectory line falling within the monitoring range of each video acquisition device can be determined by: establishing, for example... Figure 3 After creating a model of the monitoring range of all video acquisition devices corresponding to each sampling point, as shown, a trajectory line is drawn on this model to obtain the following result: Figure 4 The real-scene trajectory model is shown; traversing each sector area in the real-scene trajectory model, the length of the wearable device's trajectory line falling within the monitoring range of each video acquisition device is calculated. That is, in step S103, the historical trajectory data collected from the wearable device can be used as clues to associate it with video acquisition devices near the historical trajectory points recorded by the wearable device, and combined with the key parameters stored in the video acquisition device information database, a real-scene trajectory plane model is constructed for the trajectory traversed by the wearable device, so as to determine the degree of intersection between each video acquisition device and the historical trajectory data.

[0065] In another possible approach, the degree of intersection of the various video acquisition devices can refer to the ratio of the length of the wearable device's trajectory line falling within the monitoring range of each video acquisition device to the total length of the trajectory line.

[0066] Based on the method described above, the length of the wearable device's trajectory line falling within the monitoring range of each video acquisition device can be determined. This length can be divided by the total length of the trajectory line corresponding to the sampling point of each video acquisition device to obtain the ratio of the length of the wearable device's trajectory line falling within the monitoring range of each video acquisition device to the total length of the trajectory line. This gives the degree of intersection of each video acquisition device.

[0067] Once the degree of intersection of each video acquisition device is determined, video data from each video acquisition device can be pushed based on the degree of intersection.

[0068] Among them, the degree of intersection of video acquisition devices is positively correlated with the order in which the video acquisition devices are pushed.

[0069] The statement that "the degree of intersection between video capture devices is positively correlated with the order in which they are pushed" specifically means that when the degree of intersection between video capture devices increases, the video data from those devices tends to be pushed earlier; conversely, when the degree of intersection decreases, the video data from those devices tends to be pushed later. For example, when other variables have little or no impact on the order in which video capture devices are pushed, or when the variables affecting the order in which video capture devices are pushed remain unchanged, the higher the degree of intersection between video capture devices, the earlier the video data from those devices will be pushed.

[0070] Conversely, the lower the degree of intersection of the video capture devices, the later the video data from the video capture devices will be pushed.

[0071] Optionally, when pushing video data, video data from video acquisition devices with a zero degree of intersection may not be pushed, because the monitoring range and trajectory line of the video acquisition devices do not overlap. In this case, it is almost impossible for the target to appear in the monitoring range of the video acquisition device with a zero degree of intersection.

[0072] Within the controlled range, video data that is almost impossible for the target to appear is not pushed, reducing the amount of video data that needs to be screened and reducing video screening time. Of course, in other embodiments, video data from video acquisition devices with an intersection degree of 0 can also be pushed.

[0073] In this embodiment, firstly, historical trajectory data collected by the wearable device worn by the target is acquired, and then several sampling points in the historical trajectory data of the target's wearable device are determined.

[0074] The system searches for all video capture devices within a preset range at each sampling point. Then, based on the overlap between the monitoring range and historical trajectory data of the five searched video capture devices, it pushes video data from each device to the target location.

[0075] The degree of intersection between video capture devices is positively correlated with the order in which they are pushed to the network. Therefore, video data captured by devices with higher intersection degrees will be pushed to the network first.

[0076] The system can push the video data where the target is most likely to appear first, which can speed up the search for the target based on the pushed video data, reduce the video screening time, and thus improve the efficiency of finding the target.

[0077] Furthermore, by integrating the historical trajectory data of wearable devices with the video information from video acquisition devices, the workload of video retrieval is greatly reduced. At the same time, it eliminates the need for high-performance algorithms such as feature recognition, thereby improving retrieval efficiency and increasing the credibility of human intervention. This is crucial for tracking sensitive individuals.

[0078] Specifically, such as Figure 5 As shown, the method for finding a target proposed in this application specifically includes the following steps. It should be noted that the step numbers are for simplification only and are not intended to limit the execution order of the steps. The execution order of each step in this embodiment can be arbitrarily changed without departing from the technical concept of this application.

[0079] S201: Collect and store parameters of all video capture devices.

[0080] Optionally, parameters of all video acquisition devices within the management scope can be collected and organized. These parameters may include device model, world coordinates of installation location, orientation angle α, viewing angle β, and / or depth of field.

[0081] Among them, such as Figure 6 As shown, the orientation angle α refers to the angle between the direction the lens of the video capture device is facing and a preset direction (such as due north or due south). For devices with rotatable lenses, the initial orientation α0 during installation and the rotation angle α' during operation (this angle is transmitted back by the video capture device) can be recorded. This allows the orientation angle α corresponding to each piece of video data captured by the video capture device to be determined based on the initial orientation α0 and the rotation angle α'. For example, Figure 6 The α shown is obtained by adding the two angles mentioned above.

[0082] The viewing angle range β refers to the angle formed by the center point of the lens to the two ends of the diagonal line of the imaging plane. The viewing angle range β can be determined based on the model of the video capture equipment.

[0083] Depth of field refers to the range of distances in front of and behind a subject that can be captured in a clear image by the front edge of a video capture device. However, calculating and obtaining depth of field data requires complex calibration and measurement. In practical applications, assuming precise depth of field data cannot be obtained, since factors affecting depth of field are all related to camera parameters, the theoretical depth of field value of the video capture device, determined based on its model, can be used as the depth of field and recorded. If the depth of field of the video capture device can be calculated during the collection of its parameters, then the calculated depth of field value can be directly used as the depth of field of that video capture device.

[0084] S202: Determine several sampling points in the historical trajectory data of the wearable device of the target.

[0085] S203: Search for all video capture devices within a preset range at each sampling point.

[0086] S204: Determine the importance of each video acquisition device based on the monitoring range and the degree of intersection of historical trajectory data of each video acquisition device found.

[0087] After identifying several sampling points in the historical trajectory data of the target wearable device, the importance of each video acquisition device can be determined based on the monitoring range of each video acquisition device and the degree of intersection of the historical trajectory data. This allows the video data of each video acquisition device to be pushed in descending order of importance, thus facilitating the search for the target based on the pushed video data.

[0088] The importance of video acquisition equipment is positively correlated with the degree of overlap between the video acquisition equipment and the corresponding video acquisition equipment.

[0089] The statement that "the importance of video capture equipment is positively correlated with the degree of intersection corresponding to the video capture equipment" specifically means that as the degree of intersection corresponding to the video capture equipment increases, the importance of the video capture equipment tends to increase; conversely, as the degree of intersection corresponding to the video capture equipment decreases, the importance of the video capture equipment tends to decrease. For example, when other variables have little or no impact on the importance of the video capture equipment, or when the variables affecting the importance of the video capture equipment remain unchanged, a higher degree of intersection corresponding to the video capture equipment will increase its importance; conversely, a lower degree of intersection corresponding to the video capture equipment will decrease its importance.

[0090] In another possible implementation, the importance of the video acquisition devices can also be related to the acquisition time of the sampling points corresponding to each video acquisition device. Thus, in step S204, the importance of each video acquisition device can be determined based on the degree of intersection of each video acquisition device and the acquisition time of the sampling points corresponding to each video acquisition device.

[0091] Alternatively, the importance of the video acquisition device may be negatively correlated with the acquisition time of the sampling point corresponding to the video acquisition device.

[0092] The statement that "the importance of video capture equipment can be negatively correlated with the acquisition time of the sampling points corresponding to the video capture equipment" specifically means that if the acquisition time of the sampling points corresponding to the video capture equipment is earlier, the importance of the video capture equipment tends to decrease; conversely, if the acquisition time of the sampling points corresponding to the video capture equipment is later, the importance of the video capture equipment tends to increase. For example, when other variables have little or no impact on the importance of the video capture equipment, or when the variables affecting the importance of the video capture equipment remain unchanged, the earlier the acquisition time of the sampling points corresponding to the video capture equipment, the lower its importance; conversely, the later the acquisition time of the sampling points corresponding to the video capture equipment, the greater its importance.

[0093] In another feasible approach, the importance of video capture equipment can be determined based on other factors, such as time and location clues.

[0094] For example, if the clues include the time the target went missing, the importance of each video acquisition device at the sampling point corresponding to the time the target went missing can be increased, so that the video data of the video acquisition device corresponding to the time the target went missing can be pushed in advance, thereby reducing the video screening time and improving the efficiency of finding the target.

[0095] For example, if the clues include the address where the target went missing, the importance of each video acquisition device at the sampling point corresponding to the missing target address can be increased so that video data from the video acquisition device corresponding to the missing target address can be pushed to it in advance, thereby reducing video screening time and improving the efficiency of finding the target.

[0096] The formula for calculating the importance of video acquisition equipment is as follows:

[0097] Importance=c1×f1+c2×f2+……+c n ×f n ;

[0098] Among them, c i f is the importance coefficient of the i-th factor; i The normalized value for the importance assessment of the i-th factor, i.e., 0 ≤ f i ≤1. Optionally, the sum of the importance coefficients of all factors can be 1.

[0099] Optionally, when determining the importance of video acquisition devices, the importance of video acquisition devices with an intersection degree of 0 can be disregarded. This is because the monitoring range and trajectory line of the video acquisition devices do not overlap. In this case, the target is almost impossible to appear within the monitoring range of video acquisition devices with an intersection degree of 0. By filtering out video acquisition devices with an intersection degree of 0, video data in which the target is almost impossible to appear can be avoided, reducing the amount of video data that needs to be filtered and reducing video screening time. Of course, in other embodiments, the importance of video acquisition devices with an intersection degree of 0 can also be calculated.

[0100] S205: Push video data from each video acquisition device in descending order of importance so as to find the target based on the pushed video data.

[0101] Once all sampling points have been traversed and the importance of each video acquisition device has been determined, video data from each video acquisition device can be pushed sequentially in descending order of importance, so as to find the target based on the pushed video data.

[0102] Optionally, after recording the importance of each video acquisition device in step S204, all video acquisition devices can be sorted according to their importance, and then the video data of each video acquisition device can be pushed to the person requesting the video data in descending order.

[0103] The video data pushed from each video capture device can be video data within a certain range of capture times associated with each video capture device. That is, it is not necessary to push all video data from the video capture devices; only the video data from the video capture devices within a specified time period can be pushed.

[0104] In this implementation, historical trajectory data collected by the wearable device worn by the target is first acquired, and several sampling points in the historical trajectory data of the target's wearable device are determined. Then, all video acquisition devices within a preset range of each sampling point are searched. Subsequently, based on the monitoring range of the searched video acquisition devices and the degree of intersection with the historical trajectory data, the importance of each video acquisition device is determined. Then, the video data of each video acquisition device is pushed in descending order of importance, so as to find the target based on the pushed video data. In this way, the video data collected by the video acquisition devices with a high degree of intersection can be pushed first, so that the video data of the target most likely to appear can be pushed first, which can speed up the search for the target based on the pushed video data, reduce the video screening time, and thus improve the efficiency of finding the target. Furthermore, when determining the monitoring range of the searched video acquisition devices and the degree of intersection with historical trajectory data, a real-scene trajectory model is constructed based on historical trajectory information and the parameters of the video acquisition devices through an information fusion association algorithm. Then, the aforementioned real-scene trajectory plane model can be integrated with the historical trajectory information of wearable devices to evaluate the importance of each associated video acquisition device's real-scene video and push real-scene videos according to their importance. This greatly reduces the workload of video retrieval, reduces the use of high-performance algorithms such as feature recognition, improves retrieval efficiency, and increases the credibility of human intervention.

[0105] Furthermore, after steps S103 and S205, i.e. after the video data from the video acquisition device is pushed, key parameters such as the preset unit time, preset range, trajectory line selection time, and / or weighting coefficients of various factors in the importance calculation can be adjusted according to the pushing effect.

[0106] If the push notification is not effective, for example, if the target cannot be found based on the video data pushed, you can reduce the preset unit time, increase the preset range, increase the trajectory selection time, or adjust the weighting coefficient.

[0107] For example, after a round of the process is completed, the person who requested the video data can evaluate the video push effect, that is, manually confirm whether the pushed data meets the search requirements, that is, confirm whether the target person can be found; if the requirements are not met, key parameters such as preset unit time, preset range, trajectory line selection time and / or weighting coefficients of various factors in importance calculation can be adjusted. After the adjustment is completed, return to steps S101 and S202, re-execute the algorithm until the search task is completed.

[0108] Please see Figure 7 , Figure 7This is a schematic diagram of one embodiment of the target-finding device of this application. The target-finding device 10 includes a processor 12, which executes instructions to implement the above-described target-finding method. For detailed implementation processes, please refer to the description of the above embodiment; further details will not be repeated here.

[0109] Processor 12 can also be referred to as a CPU (Central Processing Unit). Processor 12 may be an integrated circuit chip with signal processing capabilities. Processor 12 can also be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. A general-purpose processor can be a microprocessor, or processor 12 can be any conventional processor.

[0110] The target-finding device 10 may further include a memory 11 for storing instructions and data required for the processor 12 to run.

[0111] The processor 12 is used to execute instructions to implement the methods provided by any embodiment and any non-conflicting combination of the methods for finding targets described in this application.

[0112] Please see Figure 8 , Figure 8 This is a schematic diagram of the structure of a computer-readable storage medium in an embodiment of this application. The computer-readable storage medium 30 in this embodiment stores instruction / program data 31. When executed, this instruction / program data 31 implements the method provided by any embodiment of the method for finding a target in this application, as well as any non-conflicting combination thereof. The instruction / program data 31 can be formed into a program file and stored in the storage medium 30 in the form of a software product, so that a computer device (which may be a personal computer, server, or network device, etc.) or processor can execute all or part of the steps of the methods in various embodiments of this application. The aforementioned storage medium 30 includes various media capable of storing program code, such as a USB flash drive, portable hard drive, read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk, or terminal devices such as computers, servers, mobile phones, and tablets.

[0113] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces, or indirect coupling or communication connection between apparatuses or units, and may be electrical, mechanical, or other forms.

[0114] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0115] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0116] The above are merely embodiments of this application and do not limit the scope of this patent application. Any equivalent structural or procedural changes made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the scope of patent protection of this application.

Claims

1. A method for finding a target, characterized in that, The method includes: Determine several sampling points in the historical trajectory data of the wearable device of the target; Search all video capture devices within a preset range at each sampling point; Based on the monitoring range of each video acquisition device and the degree of intersection of the historical trajectory data, video data from each video acquisition device is pushed to find the target. The order in which the video data from each video acquisition device is pushed is positively correlated with the degree of intersection of the corresponding video acquisition devices. The degree of intersection can be the ratio of the length of the wearable device's trajectory line falling within the monitoring range of each video acquisition device to the total length of the trajectory line, or the length of the wearable device's trajectory line falling within the monitoring range of each video acquisition device.

2. The method for finding a target according to claim 1, characterized in that, The process of pushing video data from each video acquisition device based on the monitoring range of each found video acquisition device and the degree of intersection between the historical trajectory data includes: Based on the monitoring range of each video acquisition device and the degree of intersection of the historical trajectory data, the importance of each video acquisition device is determined, and the importance of the video acquisition device is positively correlated with the degree of intersection of the video acquisition device. The video data from each video acquisition device is pushed sequentially in descending order of importance.

3. The method for finding a target according to claim 2, characterized in that, The determination of the importance of each video acquisition device based on the monitoring range of each video acquisition device and the degree of intersection of the historical trajectory data includes: Based on the degree of intersection of each video acquisition device and the acquisition time of the sampling point corresponding to each video acquisition device, the importance of each video acquisition device is determined. The importance of the video acquisition device is negatively correlated with the acquisition time of the sampling point corresponding to the video acquisition device.

4. The method for finding a target according to claim 1, characterized in that, The process of pushing video data from each video acquisition device based on the monitoring range of each found video acquisition device and the degree of intersection between the historical trajectory data includes: The monitoring range of each video acquisition device is determined based on the parameters of each video acquisition device. The length of the wearable device's trajectory line falling within the monitoring range of each video acquisition device is determined, and the trajectory line is determined based on the historical trajectory data; Calculate the ratio of the length corresponding to each video acquisition device to the total length of the trajectory line, and use the ratio as the degree of intersection of each video acquisition device.

5. The method for finding a target according to claim 4, characterized in that, The parameters of the video acquisition device include the coordinates, orientation angle, viewing angle, and depth of field of the video acquisition device; determining the monitoring range of each video acquisition device based on the parameters of each video acquisition device includes: Using the coordinates of each video acquisition device as the center, the depth of field of each video acquisition device as the radius, and the orientation angle and viewing angle range of each video acquisition device, the fan-shaped monitoring range of each video acquisition device is determined.

6. The method for finding a target according to claim 5, characterized in that, The search for all video acquisition devices within a preset range at each sampling point includes, prior to: Collect and store parameters of all video capture devices; If the depth of field of the video acquisition device cannot be calculated during the process of collecting parameters of the video acquisition device, the theoretical depth of field value determined based on the model of the video acquisition device will be used as the depth of field of the video acquisition device.

7. The method for finding a target according to claim 1, characterized in that, The process of pushing video data from each video acquisition device then includes: If the pushed video data cannot meet the search requirements, adjust the preset range and / or the interval between two adjacent sampling points; Based on the adjusted preset range and / or the interval between two adjacent sampling points, several sampling points from the historical trajectory data of the wearable device that performed the determination of the target are returned.

8. The method for finding a target according to claim 1, characterized in that, The determination of the target from the historical trajectory data of the wearable device includes several sampling points, including: The historical trajectory data is sampled at preset time intervals to obtain several sampling points in the historical trajectory data.

9. A device for locating a target, characterized in that, The target-finding device includes a processor; the processor is configured to execute instructions to implement the steps of the method as described in any one of claims 1-8.

10. A computer-readable storage medium storing instruction / program data thereon, characterized in that, When the instruction / program data is executed, it implements the steps of the method according to any one of claims 1-8.

Images