A target search method and system across camera devices

Abstract

The present invention relates to the field of target detection technology, and discloses a target search method and system across camera devices, which can set reasonable image acquisition frame rates for camera devices in different areas of a target location, perform target search on frame images acquired by the camera devices based on corresponding frame rates, and dynamically adjust the frame rates of camera devices in different areas based on real-time target search results, and dynamically allocate and optimize search resources in a cross-camera search scenario. Under the condition of reasonable allocation of search resources and full utilization of search resources, rapid search for preset targets can be achieved, human resources consumed by manual search can be reduced, and the search efficiency for preset targets can be improved.

Description

Technical Field

[0001] The present invention relates to the field of target detection technology, and in particular to a target search method and system across camera devices. Background Art

[0002] Hubs such as railway stations, bus stations and subway stations carry a huge amount of passengers every day and are densely populated, making it easy for people and items to be lost.

[0003] Currently, video surveillance is a key tool for finding lost persons and items. Related technologies can search for related targets through manual search, that is, staff can search for lost persons or items by viewing video surveillance.

[0004] However, manual search consumes a lot of human resources and has low search efficiency. Summary of the invention

[0005] The present invention provides a target search method and system across camera devices, which are used to solve the defects in the related art that manual search requires a large amount of human resources and has low search efficiency. The method can reduce the human resources required for manual search and improve the search efficiency for preset targets.

[0006] In a first aspect, the present invention provides a target search method across camera devices, comprising:

[0007] In response to the target search instruction, an image acquisition instruction is sent to the camera devices in the multiple sub-areas of the target location, so that the camera devices in each of the sub-areas start to acquire images based on the same first frame rate to obtain a first batch of frame images; wherein the sum of each of the first frame rates is a preset total search resource; a preset target is searched in the first batch of frame images, and it is determined that the preset target is searched in the first frame image, then the sub-area corresponding to the first frame image is determined as a main area; the multiple sub-areas are divided into a plurality of area groups according to the spatial distance between the main area and the other sub-areas; and the area groups are divided into a plurality of area groups according to the spatial distance between the area groups and the main area. The spatial distance between the area grouping and the main area is used to determine the second frame rate corresponding to the camera device in each of the area groups; wherein, the smaller the spatial distance between the area grouping and the main area, the larger the second frame rate corresponding to the camera device in the area grouping, and the sum of each of the second frame rates is the total search resource; instructing the camera device in each of the area groups to continue image acquisition based on the corresponding second frame rate to obtain a second batch of frame images; searching for the preset target in the second batch of frame images, and determining that the preset target is found in the second frame image, then determining the sub-area corresponding to the second frame image as the new main area, until a preset search end condition is met.

[0008] Optionally, the instructing the camera device in each of the area groups to continue to acquire images based on the corresponding second frame rate to obtain a second batch of frame images includes:

[0009] Instruct the camera device in each of the area groups to continue to acquire images based on the corresponding second frame rate, and divide the multiple area groups into frame rate attenuation groups, frame rate stabilization groups and frame rate increasing groups with increasing spatial distances from the main area according to the spatial distances between the area groups and the main area; instruct the frame rates of the camera devices in the frame rate attenuation groups, the frame rate stabilization groups and the frame rate increasing groups to change according to the preset frame rate attenuation strategy, frame rate stabilization strategy and frame rate increasing strategy starting from the corresponding second frame rate, and instruct each camera device to continue to acquire images according to the changed frame rate; wherein the sum of the frame rates of each camera device at any moment is not greater than the total search resource; and the frame images acquired by each camera device based on the second frame rate and the changed frame rate are taken as the second batch of frame images as a whole.

[0010] Optionally, before sending an image acquisition instruction to the camera devices in the multiple sub-areas of the target place in response to the target search instruction, the method further includes: dividing the target place into multiple functional areas according to the area function type in the target place, and taking each functional area as a sub-area; obtaining the location information of each sub-area, the connectivity information and the spatial distance between the two sub-areas; calculating the topological edge length between the two sub-areas according to the spatial distance between the two sub-areas; constructing a corresponding directed topological graph based on the location information of each sub-area, the connectivity information and the topological edge length between the two sub-areas; and determining the shortest topological path length between the two sub-areas based on the directed topological graph;

[0011] The dividing the plurality of sub-regions into a plurality of region groups according to the spatial distance between the main region and the other sub-regions comprises:

[0012] The multiple sub-areas are divided into multiple area groups according to the shortest topological path length between the main area and other sub-areas.

[0013] Optionally, dividing the multiple sub-areas into multiple area groups according to the topological path lengths between the main area and the other sub-areas includes:

[0014] Among the shortest topological path lengths between the sub-areas, the maximum length is determined; a first product obtained by multiplying the maximum length by a preset first proportional coefficient is calculated, and a second product obtained by multiplying the maximum length by a preset second proportional coefficient is calculated; wherein the first proportional coefficient and the second proportional coefficient are greater than 0 and less than 1, and the first proportional coefficient is less than the second proportional coefficient; among the multiple sub-areas, a first sub-area whose shortest topological path length with the main area is less than the first product is determined, a second sub-area whose shortest topological path length with the main area is greater than the first product and less than the second product is determined, and a third sub-area whose shortest topological path length with the main area is greater than the second product is determined; all the first sub-areas and the main area as a whole are grouped as a first area, all the second sub-areas as a whole are grouped as a second area, all the third sub-areas as a whole are grouped as a third area, and the first area grouping, the second area grouping and the third area grouping are grouped as the multiple area groups.

[0015] Optionally, dividing the plurality of regional groups into frame rate attenuation groups, frame rate stabilization groups and frame rate increasing groups whose spatial distances from the main region increase in sequence according to the spatial distances between the regional groups and the main region comprises:

[0016] The first area group, the second area group and the third area group are respectively used as the frame rate decay group, the frame rate stabilization group and the frame rate increasing group.

[0017] Optionally, the total search resource is the maximum number of frame images that can be processed within a unit time when performing target search based on frame images;

[0018] The first frame rate is obtained by dividing the maximum number of frame images by the total number of cameras in the target location; the second frame rate corresponding to the cameras in the first area grouping is a preset maximum allowable frame rate; the second frame rate corresponding to the cameras in the second area grouping is equal to the first frame rate; the second frame rate corresponding to the cameras in the third area grouping is a preset minimum allowable frame rate; wherein, the first frame rate is greater than the minimum allowable frame rate, and less than the maximum allowable frame rate.

[0019] Optionally, the respectively instructing the frame rates of the camera devices in the frame rate decay group, the frame rate stabilization group, and the frame rate increasing group to change according to a preset frame rate decay strategy, frame rate stabilization strategy, and frame rate increasing strategy starting from the corresponding second frame rate, includes:

[0020] Determine the first average flow of people of a target group in the preset past number of days, the target group being the frame rate attenuation group, the frame rate stability group or the frame rate increasing group; for any sub-area in the target group, determine the second average flow of people of the sub-area in the preset past number of days, divide the second average flow of people by the first average flow of people to obtain the target appearance probability of the sub-area, construct a frame rate change expression according to the target appearance probability of the sub-area, and instruct the frame rate of the camera device in the sub-area to change according to the frame rate change expression starting from the corresponding second frame rate.

[0021] Optionally, when the target group is the frame rate attenuation group, the frame rate change expression constructed according to the target appearance probability of the sub-area in the target group is used to decay the frame rate of the camera device in the sub-area from the corresponding second frame rate to the first frame rate; when the target group is the frame rate increasing group, the frame rate change expression constructed according to the target appearance probability of the sub-area in the target group is used to increase the frame rate of the camera device in the sub-area from the corresponding second frame rate to the first frame rate.

[0022] Optionally, when the target group is the frame rate stable group, constructing a frame rate change expression according to the target appearance probability of the sub-area includes:

[0023] Multiplying the target occurrence probability of the sub-region by the first frame rate to obtain a target product, and using the target product as the frame rate change expression;

[0024] The instructing the frame rate of the camera device in the sub-area to change from the corresponding second frame rate according to the frame rate change expression includes:

[0025] The frame rate of the camera in the sub-area is instructed to be adjusted from the corresponding second frame rate and stabilized to the target product.

[0026] In a second aspect, the present invention provides a target search system across camera devices, comprising:

[0027] A sending unit, configured to send an image acquisition instruction to the camera devices in the plurality of sub-areas of the target location in response to the target search instruction, so that the camera devices in each of the sub-areas start to acquire images based on the same first frame rate to obtain the first batch of frame images; wherein the sum of each of the first frame rates is a preset total search resource;

[0028] A first search unit, used for searching for a preset target in the first batch of frame images;

[0029] A first determining unit, configured to determine that the preset target is searched in the first frame image, and then determine the sub-region corresponding to the first frame image as the main region;

[0030] a division unit, configured to divide the plurality of sub-regions into a plurality of region groups according to the spatial distance between the main region and the other sub-regions;

[0031] A second determination unit is used to determine a second frame rate corresponding to the camera device in each of the area groups according to a spatial distance between the area group and the main area; wherein the smaller the spatial distance between the area group and the main area, the larger the second frame rate corresponding to the camera device in the area group, and the sum of each of the second frame rates is the total search resource;

[0032] An instruction unit, used for instructing the camera device in each of the area groups to continue to collect images based on the corresponding second frame rate to obtain a second batch of frame images;

[0033] A second search unit, configured to search for the preset target in the second batch of frame images;

[0034] The third determining unit is used to determine that if the preset target is found in the second frame image, the sub-region corresponding to the second frame image is determined as a new main region until a preset search end condition is met.

[0035] The cross-camera target search method and system provided by the present invention can set reasonable image acquisition frame rates for cameras in different areas of a target location, perform target search on frame images acquired by the cameras based on the corresponding frame rates, and dynamically adjust the frame rates of cameras in different areas based on real-time target search results. In a cross-camera search scenario, dynamic allocation and allocation optimization of search resources can be performed, and rapid search for preset targets can be achieved while reasonably allocating search resources and fully utilizing search resources, thereby reducing the human resources consumed by manual search and improving the search efficiency for preset targets. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0037] Figure 1 A flow chart of a target search method across camera devices provided by an embodiment of the present invention;

[0038] Figure 2A directed topological graph corresponding to a target location provided in an embodiment of the present invention;

[0039] Figure 3 An adjacency matrix provided by an embodiment of the present invention;

[0040] Figure 4 A shortest path matrix provided by an embodiment of the present invention;

[0041] Figure 5 A schematic diagram of area grouping provided by an embodiment of the present invention;

[0042] Figure 6 A flowchart of another target search method across camera devices provided by an embodiment of the present invention;

[0043] Figure 7 A flowchart of another target search method across camera devices provided by an embodiment of the present invention;

[0044] Figure 8 A schematic diagram of the structure of a target search system across camera devices provided by an embodiment of the present invention;

[0045] Fig. 9 A schematic diagram of the structure of a computer device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0046] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0047] In order to solve the problem that manual search in related technologies consumes a lot of human resources and has low search efficiency, the present embodiment proposes a target search method and system across camera devices, which can set reasonable image acquisition frame rates for camera devices in different areas of the target location, perform target search on frame images acquired by the camera devices based on the corresponding frame rates, and dynamically adjust the frame rates of camera devices in different areas based on real-time target search results. In the cross-camera search scenario, dynamic allocation and allocation optimization of search resources are performed, which can achieve rapid search of preset targets while reasonably allocating search resources and fully utilizing search resources, reduce human resources consumed by manual search, and improve search efficiency for preset targets.

[0048] Combine the following Figure 1-Figure 7 A target search method across camera devices according to an embodiment of the present invention is described.

[0049] like Figure 1 As shown, this embodiment proposes a first target search method across camera devices, which may include the following steps:

[0050] S101. In response to a target search instruction, an image acquisition instruction is sent to the camera devices in multiple sub-areas of the target location, so that the camera devices in each sub-area start image acquisition based on the same first frame rate to obtain the first batch of frame images; wherein the sum of each first frame rate is a preset total search resource.

[0051] It should be noted that this embodiment can be applied to a server. This embodiment can search for relevant targets in a frame image based on a target search algorithm.

[0052] Specifically, the target search instruction can be input into the server by the staff. Specifically, the staff can input the target search instruction into the server when they need to search for the target in the target place.

[0053] Among them, the target places can be railway stations, passenger stations, subway stations, shopping malls, hospitals, schools and other places.

[0054] Specifically, the sub-area is a partial area in the target place. A camera device may be arranged in each sub-area. It should be noted that one or more cameras may be arranged in each sub-area. The camera device captures images of the space in the sub-area where it is located.

[0055] Specifically, the camera device is a device used for performing video surveillance in a corresponding area, such as a camera or a webcam.

[0056] The total search resource may be the total number of frame images that the server can process at most within a unit time.

[0057] It can be understood that the first frame rate can be obtained by dividing the total search resources by the total number of camera devices in the above-mentioned multiple sub-areas.

[0058] Specifically, when the server receives a target search instruction, it can respond to the target search instruction and send an image acquisition instruction to the camera device in each sub-area, so that the camera device in each sub-area performs image acquisition based on the first frame rate. At this time, the server will evenly distribute the total search resources to each camera device.

[0059] The first batch of frame images may include all frame images captured by all camera devices in the above-mentioned multiple sub-areas when performing image capture based on the first frame rate.

[0060] S102: Searching for a preset target in the first batch of frame images.

[0061] Among them, the preset target is the target that needs to be searched.

[0062] Optionally, the preset target can be a certain person or an object.

[0063] Specifically, this embodiment may search for a preset target in each frame image of the first batch of frame images based on a target search algorithm.

[0064] S103: If it is determined that the preset target is found in the first frame image, the sub-region corresponding to the first frame image is determined as the main region.

[0065] It should be noted that hubs such as railway stations, passenger stations and subway stations carry huge passenger flow and dense crowds every day, which makes it very easy for people and items to be lost. Video surveillance is the key medium and tool for finding lost people and items. By superimposing the target detection algorithm on the video surveillance of the station camera, it is possible to quickly find the lost target person and target item, improve the search efficiency of the staff, and reduce the workload of the staff. In the related art, the computing resources occupied by the target detection and recognition algorithm are large, and there are a large number of monitors in the station. Due to the limitation of the computing resources of the station server, it is impossible to ensure that the target detection algorithm is enabled for all the images of all the monitoring devices in the station at all times. Therefore, it is necessary to collect and extract the camera images at regular intervals, and then input the extracted images into the target detection algorithm for detection. If the computing resources of the camera cannot be reasonably allocated and optimized, it will affect the quality and efficiency of the camera's search for lost people and items, resulting in the camera's inability to quickly find the lost people and items.

[0066] In order to make full use of the limited computing resources of the camera server and improve the efficiency of searching for lost persons and objects, this implementation proposes a dynamic resource allocation strategy in a cross-camera search scenario. By setting a reasonable acquisition frame rate for the camera device in the search area of ​​the target location, the captured frame images are used for target search, thereby achieving rapid detection and search for lost persons and objects, and finding the movement trajectory of lost persons or objects based on time and space, thereby achieving accurate positioning of lost persons and objects.

[0067] The first frame image may be a frame image in the first batch of frame images.

[0068] Specifically, the sub-region corresponding to the first frame image is the sub-region where the camera device that captures the first frame image is located. It can be understood that the camera device captures the first frame image during the image capture process for the space of the sub-region.

[0069] It should be noted that, in this embodiment, when the preset target is searched in the first frame image, the frame rate allocation strategy for the camera device in the target location can be adjusted. The general allocation principle is that the closer the distance to the preset target area is, the more frame rate is allocated, and the farther the distance to the preset target area is, the less frame rate is allocated.

[0070] S104: Divide the multiple sub-regions into multiple region groups according to the spatial distances between the main region and other sub-regions.

[0071] Among them, when determining the spatial distance between two sub-regions, this embodiment can first determine the position points in the two sub-regions that best represent the positions of the regions, such as the geometric center points, centroid points or symmetry points of the sub-regions, and then calculate the distance between the position points in the two sub-regions that best represent the positions of the regions, which is used as the spatial distance between the two sub-regions.

[0072] Specifically, this embodiment can first determine the spatial distances between the main area and each other sub-area, then determine multiple distance thresholds based on the spatial distances between the main area and each other sub-area, and then divide the multiple sub-areas into multiple area groups based on the multiple distance thresholds.

[0073] For example, this embodiment can determine the maximum and minimum values ​​in the spatial distances between the main area and each of the other sub-areas, determine multiple distance thresholds in the interval from the maximum value to the minimum value (such as dividing the interval into multiple sub-intervals evenly, and determining the initial values ​​and end values ​​of these sub-intervals that are greater than the minimum value and less than the maximum value as distance thresholds), and divide the multiple sub-areas into multiple area groups according to the multiple distance thresholds.

[0074] For another example, the present embodiment may first determine the number n of area groups to be divided, determine the maximum value in the spatial distances between the main area and each of the other sub-areas, and then multiply the maximum value by one-n, two-n, and so on, by m-n, where m is n minus 1. At this time, m products can be obtained and used as distance thresholds, that is, m distance thresholds are obtained, and multiple sub-areas are divided into n area groups according to the m distance thresholds.

[0075] It can be understood that, in this embodiment, multiple distance thresholds may also be determined based on the minimum value, average value, or other characteristic distances among the spatial distances between the main region and each of the other sub-regions.

[0076] It should be noted that each area group may include one or more sub-areas, and the sub-areas in different area groups are different.

[0077] It should also be noted that, after determining multiple distance thresholds, this embodiment can sort the multiple distance thresholds, and then divide the multiple sub-regions into multiple region groups according to the sorted distance thresholds. For example, when the sorted distance thresholds include three distance thresholds A, B and C that increase in sequence, this embodiment can group all sub-regions whose spatial distance from the main region is less than or equal to A as a whole region, all sub-regions whose spatial distance from the main region is greater than A and less than or equal to B as a whole region, all sub-regions whose spatial distance from the main region is greater than B and less than or equal to C as a whole region, and all sub-regions whose spatial distance from the main region is greater than C as a whole region.

[0078] S105. Determine the second frame rate corresponding to the camera device in each area group according to the spatial distance between the area group and the main area; wherein, the smaller the spatial distance between the area group and the main area, the larger the second frame rate corresponding to the camera device in the area group, and the sum of each second frame rate is the total search resource.

[0079] Specifically, when determining the spatial distance between the regional grouping and the main area, this embodiment can first determine the position points in the regional grouping and the main area that best represent the location of the area, such as the geometric center point, the centroid point or the symmetry point, and then calculate the distance between the position points in the regional grouping and the main area that best represent the location of the area, which is used as the spatial distance between the regional grouping and the main area.

[0080] Specifically, this embodiment can allocate the second frame rate to the camera devices in different regional groups according to the spatial distance between the regional group and the main area. This embodiment can set the frame rate allocation principle as follows: the smaller the spatial distance between the regional group and the main area, the larger the second frame rate corresponding to the camera device in the regional group, the second frame rates corresponding to the camera devices in the same regional group are the same, and the sum of each second frame rate is the total search resource.

[0081] Optionally, this embodiment may first sort the spatial distances between all regional groups and the main area, determine the spatial distance in the middle according to the sorted spatial distances, and use it as the target spatial distance, and use the regional group corresponding to the target spatial distance as the target regional group. After that, this embodiment may directly use the first frame rate as the second frame rate and assign it to each camera device in the target regional group, that is, the frame rate of each camera device in the target regional group remains unchanged. This embodiment may further set the frame rate allocation principle as follows: for the regional group whose spatial distance from the main area is less than the target spatial distance, determine that the second frame rate assigned to each camera device in the regional group is greater than the first frame rate; for the regional group whose spatial distance from the main area is greater than the target spatial distance, determine that the second frame rate assigned to each camera device in the regional group is less than the first frame rate. It should be noted that the smaller the spatial distance between the regional group and the main area, the larger the second frame rate corresponding to the camera device in the regional group. The second frame rates of the cameras in the same regional group are the same.

[0082] At this time, this embodiment can establish an equation related to the total search resources based on the server's total search resources, the number of cameras in each area grouping, and the frame rate allocation principle for cameras in different area groups, solve the allowable frame rate range of cameras in different area groups, and then select a suitable value from the allowable frame rate range of cameras in the area grouping as the second frame rate of the camera.

[0083] Optionally, the present embodiment may also pre-set the maximum allowable frame rate and the minimum allowable frame rate of the camera device, and allocate the maximum allowable frame rate as the second frame rate to each camera device in the regional group closest to the main area, and allocate the minimum allowable frame rate as the second frame rate to each camera device in the regional group farthest from the main area, and then establish an equation related to the total search resources based on the allocated search resources, the spatial distance between other regional groups and the main area, and the number of cameras in each regional group, solve the allowable frame rate range of the camera devices in different regional groups, and then select a suitable value from the allowable frame rate range of the camera devices in the regional group as the second frame rate of the camera device.

[0084] S106 , instructing the camera device in each area group to continue to collect images based on the corresponding second frame rate to obtain a second batch of frame images.

[0085] In this embodiment, the camera device in each area group may be instructed to adjust from the first frame rate to the corresponding second frame rate and continue to capture images.

[0086] Specifically, the second batch of frame images may include all frame images captured by each camera device during the image capture process based on the corresponding second frame rate.

[0087] S107: Searching for a preset target in the second batch of frame images.

[0088] Specifically, this embodiment may search for a preset target in each frame image of the second batch of frame images based on a target search algorithm.

[0089] S108: If it is determined that the preset target is found in the second frame image, the sub-region corresponding to the second frame image is determined as a new main region until a preset search end condition is met.

[0090] The second frame image is a frame image in the second batch of frame images.

[0091] Specifically, this embodiment can determine the camera device that captures the second frame image, and determine the sub-area where the camera device is located as the new main area, adjust the frame rate of the camera devices in each sub-area according to the new main area, and instruct each camera device to continue image capture based on the adjusted frame rate until the preset search end condition is met.

[0092] Specifically, the search end condition may be that the total search time reaches a preset time threshold, or that the staff finds the preset target offline. The total search time may be the time obtained by counting from the moment when step S101 is executed in this embodiment.

[0093] Optionally, the search end condition may also be to determine that the preset target has moved to an area outside the target location. Specifically, in this embodiment, when all the camera devices in the target location do not detect the preset target within a certain period of time, that is, in this embodiment, when all the camera devices in the target location do not detect the preset target in the frame images collected within a certain continuous time, it is determined that the preset target has moved outside the station. The specific time can be determined and preset according to the actual situation.

[0094] It should be noted that the target search algorithm detects the target only to indicate when the target appeared in a certain place. In order to assist the staff to actually find the target offline, the present embodiment can record the time and location of the preset target and provide it to the staff, which helps the staff to determine the main area and search route for offline search, and facilitates the staff to find the target person or item at the first time. For example, the main area can be determined as the main area, and the staff can search for the preset target offline according to the determined main area and search route. When the staff searches for the preset target offline, the present embodiment can end the current search task.

[0095] There are a large number of cameras in target places such as passenger stations, railway stations and subway stations, and the video surveillance system is relatively developed. However, the algorithm server resources supporting the target detection of the cameras are limited. The target detection algorithm can only be deployed for the frame images collected by a fixed number of cameras within a certain period of time for target detection. There is a lack of reasonable allocation of limited resources, and it is easy to miss the best time window for finding people or objects, resulting in more waste of computing power and time resources, and even causing the problem of target search failure. In order to make full use of the limited camera computing resources and improve the efficiency of the camera in finding lost people and objects, this embodiment proposes a dynamic resource allocation strategy in a cross-camera search scenario. By setting different image frame acquisition strategies for cameras in different areas, that is, all cameras in the station are divided into several areas, and different video frame acquisition strategies are set for cameras in different areas, rapid retrieval of people and target objects is achieved, and the time and route of the search target can be recorded, so that the staff can find the target person and object offline as soon as possible.

[0096] The cross-camera target search method proposed in this embodiment can set reasonable image acquisition frame rates for cameras in different areas of the target location, perform target search on frame images acquired by the cameras based on the corresponding frame rates, and dynamically adjust the frame rates of cameras in different areas based on real-time target search results. In a cross-camera search scenario, dynamic allocation and optimization of search resources can be performed, and rapid search for preset targets can be achieved while reasonably allocating search resources and fully utilizing them, thereby reducing the human resources consumed by manual search and improving the search efficiency for preset targets.

[0097] based on Figure 1 This embodiment proposes a second target search method across camera devices. In this method, the above step S106 may include:

[0098] Instructing the camera device in each area group to continue to acquire images based on the corresponding second frame rate, and dividing the multiple area groups into a frame rate decaying group, a frame rate stable group, and a frame rate increasing group, whose spatial distances from the main area increase in sequence, according to the spatial distances between the area groups and the main area;

[0099] Instruct the frame rates of the camera devices in the frame rate decay group, the frame rate stabilization group, and the frame rate increment group to change from the corresponding second frame rate according to the preset frame rate decay strategy, frame rate stabilization strategy, and frame rate increment strategy, and instruct each camera device to continue to collect images according to the changed frame rate; wherein the sum of the frame rates of each camera device at any time is not greater than the total search resource;

[0100] The frame images collected by each camera device based on the second frame rate and the changed frame rate are taken as a second batch of frame images.

[0101] It should be noted that, due to the limited search resources of the server and the characteristics of the preset target moving over time, this embodiment can set a frame rate change strategy for the camera devices in different areas. In this embodiment, for a type of area close to the main area, the search target has a high probability of appearing in a short period of time, so the camera devices in this type of area can be set to a higher frame rate at the beginning, and as time goes by, when the main area cannot detect the preset search target, the frame rate of this type of area should be gradually reduced to save computing resources.

[0102] For a type of area that is far from the main area, the probability of the search target appearing in a short period of time is small, so the camera device in this type of area can be set to a smaller frame rate at the beginning. As time goes by, when the main area cannot detect the preset search target, the frame rate of this type of area can be gradually increased to make full use of computing resources, maintain total resources and increase the possibility of retrieving the target. For a type of area that is a moderate distance from the main area, in order to ensure resource utilization and search target retrieval efficiency, the acquisition frame rate of the camera device in this type of area can remain unchanged in a short period of time.

[0103] Specifically, in this embodiment, the above-mentioned multiple regional groups can be further divided into frame rate decay groups, frame rate stability groups and frame rate increasing groups according to the spatial distance between the regional groups and the main area. Among them, the frame rate decay groups, frame rate stability groups and frame rate increasing groups can all include one or more regional groups, and the frame rate decay groups, frame rate stability groups and frame rate increasing groups contain different regional groups.

[0104] It can be understood that this embodiment can refer to the specific method and process of dividing multiple sub-areas into multiple area groups according to the spatial distance between the main area and other sub-areas in this embodiment, and further divide the multiple area groups into frame rate attenuation groups, frame rate stable groups and frame rate increasing groups.

[0105] Specifically, the spatial distance between the frame rate decaying group and the main area is greater than the spatial distance between the frame rate stable group and the main area, and the spatial distance between the frame rate stable group and the main area is greater than the spatial distance between the frame rate increasing group and the main area.

[0106] Specifically, this embodiment can adjust the frame rate of each camera device in the frame rate attenuation group from the corresponding second frame rate according to the preset frame rate attenuation strategy, can adjust the frame rate of each camera device in the frame rate stabilization group from the corresponding second frame rate according to the preset frame rate stabilization strategy, and can adjust the frame rate of each camera device in the frame rate increasing group from the corresponding second frame rate according to the preset frame rate increasing strategy.

[0107] The sum of the frame rates of all the cameras in the target location at any time is not greater than the total search resource.

[0108] Specifically, in this embodiment, all frame images acquired by all camera devices in the target location based on the corresponding second frame rate and the frame rate changed from the second frame rate based on the corresponding strategy can be taken as a whole as the second batch of frame images.

[0109] The target search method across camera devices proposed in this embodiment can further divide the above-mentioned multiple area groups into frame rate attenuation groups, frame rate stability groups and frame rate increasing groups according to the spatial distance between them and the main area, and adopt different frame rate change strategies for the camera devices in the frame rate attenuation groups, frame rate stability groups and frame rate increasing groups, so as to further realize the dynamic allocation and optimized allocation of server search resources, and further realize the reasonable allocation and full utilization of search resources.

[0110] Based on the second target search method across camera devices, this embodiment proposes a third target search method across camera devices. The third target search method across camera devices may further include, before step S101:

[0111] According to the regional functional type in the target place, the target place is divided into multiple functional areas, and each functional area is regarded as a sub-area;

[0112] Obtain the location information of each sub-region, the connectivity information between two sub-regions, and the spatial distance;

[0113] According to the spatial distance between two sub-regions, the topological edge length between two sub-regions is calculated;

[0114] Based on the location information of each sub-region, the connectivity information between two sub-regions and the topological edge length, a corresponding directed topological graph is constructed;

[0115] Based on the directed topological graph, the shortest topological path length between any two sub-areas is determined.

[0116] At this time, the above step S104 may include:

[0117] According to the shortest topological path length between the main area and other sub-areas, multiple sub-areas are divided into multiple area groups.

[0118] Specifically, in this embodiment, the target place can be divided into a plurality of functional areas according to the functions of different areas in the target place, and each functional area is a sub-area.

[0119] Among them, the location information of the sub-area can be used to identify the location of the sub-area in the target place, such as the coordinates of the sub-area in the coordinate system of the target place. The connectivity information between sub-areas can identify whether there is a passable path between the sub-areas, and determine whether the passable path is a one-way path or a two-way path. It should be noted that when two sub-areas can be passed in both directions in an actual scene, the passable path between the two sub-areas is a two-way path, and when the two sub-areas can only be passed in one direction in an actual scene, the passable path between the two sub-areas is a one-way path.

[0120] Specifically, this embodiment can create a corresponding topological node according to each sub-region, and convert the spatial distance between two sub-regions into the corresponding topological edge length. For example, this embodiment can divide the spatial distance between sub-regions by 100 to obtain the corresponding topological edge length. For example, when the spatial distance between two sub-regions is 50 meters, the corresponding topological edge length is 0.5.

[0121] Specifically, this embodiment can construct a corresponding directed topological graph based on the connectivity information between all sub-areas, all topological nodes and corresponding topological edge lengths.

[0122] In order to better introduce the construction process of the directed topological graph, this embodiment takes a certain place in a railway station as the target place for explanation. In this embodiment, according to the eight functions of the railway station, the railway station can be divided into eight sub-areas, namely, the corridor, the wide hall, the second waiting room, the waiting room for special groups, the third waiting room, the fifth waiting room, the shopping gallery and the fourth waiting room. In the eight sub-areas, every two sub-areas are bidirectionally connected.

[0123] like Figure 2 As shown, this embodiment can create topological nodes numbered 1 to 8 based on the 8 divided sub-areas, which are used to identify the 8 sub-areas respectively, convert the spatial distances between any two sub-areas into corresponding topological edge lengths, and construct a corresponding directed topological graph based on the location information of the 8 sub-areas, the connectivity information between the 8 sub-areas, the created 8 topological nodes and the corresponding topological edge lengths. Figure 2 As shown in the figure, the line between two topological nodes is the edge between the topological nodes. The bidirectional arrow of the line indicates bidirectional connectivity. The lengths of the topological edges are 0.5, 0.8, 1, and 2.

[0124] Specifically, this embodiment can calculate the adjacency matrix based on the constructed directed topological graph. The adjacency matrix is ​​a matrix that represents the adjacent relationship between the vertices of the graph. Let G = (V, E) be a graph, where V is the vertex and E is the edge. Therefore, a one-dimensional array can be used to store the data of all vertices in the graph; a two-dimensional array can be used to store the data of the relationship between vertices (edges or arcs). The adjacency matrix can be a directed graph adjacency matrix and an undirected graph adjacency matrix. In this embodiment, the adjacency matrix of the directed graph is used to represent the degree of association between topological nodes. A[ g,p ] represents a two-dimensional array of node topological relationships. This embodiment can construct a topological edge length between two topological nodes as follows: Figure 3 The adjacency matrix A shown in FIG. 1 records the topological edge lengths between each pair of topological nodes.

[0125] Specifically, this embodiment may calculate the shortest path length between any two topological nodes, that is, the shortest topological path length between any two sub-areas, based on the shortest path algorithm.

[0126] Optionally, this embodiment can use Dijkstra algorithm or other shortest path algorithms to calculate the shortest path and the shortest path length between two topological nodes, and obtain the shortest path matrix corresponding to the adjacency matrix A, which includes the shortest topological path length between two sub-areas. Dijkstra When the algorithm calculates the shortest path matrix corresponding to the adjacency matrix A, the following process can be performed:

[0127] (1) Set two vertex sets U and T. Set U stores the vertices whose shortest paths have been found, and set T stores the vertices whose shortest paths have not yet been found.

[0128] (2) In the initial state, the set U only contains the source point, which is set to v0;

[0129] (3) Then select the vertex u with the shortest path length to the source point v0 from the set T and add it to the set U;

[0130] (4) Every time a new vertex u is added to the set U, the current shortest path value between the source point v0 and the remaining vertices in the set T must be modified. The new current shortest path length value of each vertex in the set T is the smaller of the original current shortest path length value and the weighted path length from the source point through vertex u to the vertex.

[0131] (5) Return to step (3) and repeat this step until all vertices in set T are added to set U.

[0132] At this time, this embodiment can obtain Figure 4 The shortest path matrix shown.

[0133] Optionally, the multiple sub-areas are divided into multiple area groups according to the topological path lengths between the main area and other sub-areas, including:

[0134] Among the shortest topological path lengths between two sub-areas, determine the maximum length;

[0135] Calculating a first product obtained by multiplying the maximum length by a preset first proportional coefficient, and calculating a second product obtained by multiplying the maximum length by a preset second proportional coefficient; wherein the first proportional coefficient and the second proportional coefficient are greater than 0 and less than 1, and the first proportional coefficient is less than the second proportional coefficient;

[0136] Among the multiple sub-areas, determine a first sub-area whose shortest topological path length with the main area is less than the first product, determine a second sub-area whose shortest topological path length with the main area is greater than the first product and less than the second product, and determine a third sub-area whose shortest topological path length with the main area is greater than the second product;

[0137] All first sub-regions and the main region as a whole are grouped as a first region, all second sub-regions as a whole are grouped as a second region, all third sub-regions as a whole are grouped as a third region, and the first region group, the second region group and the third region group are grouped as a plurality of region groups.

[0138] Specifically, in this embodiment, after creating multiple topological nodes, the attribute characteristics of each topological node can be set. r The attribute characteristics include (x r ,y r ,N r ,C r (P r ,t)). Among them, r is the serial number of the topological node, V r is a topological node with serial number r, (x r ,y r ) represents the topological node V r The position coordinates, N r represents the number of cameras in the sub-area corresponding to the topological node, C r (P r ,t) represents the topological node V r The image frame acquisition strategy of the camera device, that is, how many image frames the camera device acquires per second, P r Represents the probability of the preset target appearing in the sub-area corresponding to the topological node. The total search resource of the server is the total number of frame images that the server can process per second. , the number of cameras in the target location is , then the average acquisition frame rate of each camera device when maximizing the use of computing resources is for Divide by It can be understood that the first frame rate is , the number of cameras in the target location That is, the sum of the number of cameras in each sub-area.

[0139] Specifically, when the computing service center of the target location has not arranged a search task, this embodiment can prohibit the camera device in each sub-area from performing image acquisition. When receiving a search task, i.e., a target search instruction, this embodiment can set the same frame image acquisition strategy for the camera device in each sub-area, that is, set the frame rate of each camera device to the first frame rate. In this embodiment, when a preset target is searched in a frame image captured by a camera device in a sub-area, the sub-area can be set as the main area, and the time when the preset target is searched is recorded as Ts. At this time, the frame rate of the camera device in the main area can be set to the maximum allowable frame rate. , record the movement of the searched target in the place as much as possible. At time Ts, this embodiment can divide the shortest distance between each topological node in the directed topological graph and the main node (i.e., the topological node corresponding to the main area) into three categories according to the shortest path matrix. The node set closest to the main node The frame rate of the camera device, that is, the initial video frame acquisition strategy, is set to , the set of nodes farthest from the master node The frame rate of the camera device, that is, the initial video frame acquisition strategy, is set to , the last type of node set That is, the frame rate of the camera device in a class of nodes that are moderately far from the main node, that is, the initial video frame acquisition strategy is set to .

[0140] The first proportionality coefficient and the second proportionality coefficient can be set by a technician according to actual conditions. For example, the first proportionality coefficient and the second proportionality coefficient can be set to one third and two thirds respectively.

[0141] Specifically, in this embodiment, the maximum length can be determined from the shortest topological path lengths between two sub-regions, and then the first product and the second product are calculated based on the maximum length, the first proportional coefficient and the second proportional coefficient, and then the first region group, the second region group and the third region group are determined according to the first product and the second product. In this embodiment, the frame rate of each camera device in the first region group can be set to , set the frame rate of each camera device in the second area group to , set the frame rate of each camera device in the third area group to .in, and The value can be obtained by the following formula:

[0142] .

[0143] in, Represents topological node V r The shortest topological path length between the corresponding sub-area and the main area, i is the sub-area number corresponding to the node number, Represents the maximum length of the shortest topological path between any two sub-areas. For the node class in the directed topology graph The number of included camera devices, that is, the number of camera devices in the first area group; For the node class in the directed topology graph The number of included camera devices, that is, the number of camera devices in the third area group; For the node class in the directed topology graph The number of included camera devices, that is, the number of camera devices in the second area group. The total search resources of the server are the total number of frame images that the server can process per second. is the number of camera devices in the target location, and one third and two thirds represent the first proportional coefficient and the second proportional coefficient respectively. and are the first product and the second product respectively.

[0144] based on Figure 2 The directed topological graph corresponding to the railway station shown in the figure can be obtained after processing in this embodiment. Figure 5 The regional grouping diagram is shown in Figure 1. Figure 5 In the figure, topological node 2 is the main node, that is, the sub-area hall corresponding to topological node 2 is the main area. The first area group includes the sub-areas corresponding to topological nodes 2, 1, 3, 4, 5 and 6. The second area group includes the sub-area corresponding to topological node 7. The third area group includes the sub-area corresponding to topological node 8.

[0145] Optionally, the above-mentioned dividing the multiple regional groups into frame rate attenuation groups, frame rate stability groups and frame rate increasing groups whose spatial distances from the main area increase in sequence according to the spatial distances between the regional groups and the main area includes:

[0146] The first area group, the second area group and the third area group are respectively used as a frame rate decay group, a frame rate stabilization group and a frame rate increasing group.

[0147] It can be understood that the spatial distance between the first area group and the main area is smaller than the spatial distance between the second area group and the main area. The spatial distance between the second area group and the main area is smaller than the spatial distance between the third area group and the main area. Therefore, the spatial distances between the first area group, the second area group and the third area group and the main area increase in sequence.

[0148] Specifically, when the number of area groups is 3, this embodiment can directly determine the first area group, the second area group, and the third area group as a frame rate attenuation group, a frame rate stabilization group, and a frame rate increasing group whose spatial distances from the main area increase in sequence. That is, this embodiment can directly determine the first area group as a frame rate attenuation group, the second area group as a frame rate stabilization group, and the third area group as a frame rate increasing group.

[0149] Optionally, the total search resource is the maximum number of frame images that can be processed within a unit time when performing target search based on frame images;

[0150] The first frame rate is obtained by dividing the maximum number of frame images by the total number of camera devices in the target location;

[0151] The second frame rate corresponding to the camera device in the first area group is the preset maximum allowable frame rate; the second frame rate corresponding to the camera device in the second area group is equal to the first frame rate; the second frame rate corresponding to the camera device in the third area group is the preset minimum allowable frame rate;

[0152] The first frame rate is greater than a minimum allowed frame rate and less than a maximum allowed frame rate.

[0153] Optionally, the frame rates of the camera devices in the frame rate decay group, the frame rate stabilization group, and the frame rate increasing group are respectively instructed to change starting from the corresponding second frame rate according to a preset frame rate decay strategy, frame rate stabilization strategy, and frame rate increasing strategy, including:

[0154] Determine the first average flow of people of the target group in the preset past days, the target group being a frame rate decaying group, a frame rate stable group or a frame rate increasing group;

[0155] For any sub-area in the target group, the second average flow of people in the sub-area in the preset past number of days is divided by the first average flow of people to obtain the target appearance probability of the sub-area, and a frame rate change expression is constructed according to the target appearance probability of the sub-area. The frame rate of the camera device in the sub-area is instructed to change according to the frame rate change expression starting from the corresponding second frame rate.

[0156] The first average flow of people may be the average flow of people Q1 of the target group in the past preset number of days, and the second average flow of people may be the average flow of people Q2 of a sub-area in the target group in the past preset number of days. In this embodiment, Q2 may be divided by Q1, and the obtained value may be used as the probability of target appearance.

[0157] Optionally, when the target group is a frame rate attenuation group, a frame rate change expression constructed according to the target appearance probability of a sub-region in the target group is used to attenuate the frame rate of the camera device in the sub-region from the corresponding second frame rate to the first frame rate;

[0158] When the target group is a frame rate increasing group, the frame rate change expression constructed according to the target appearance probability of the sub-area in the target group is used to increase the frame rate of the camera device in the sub-area from the corresponding second frame rate to the first frame rate.

[0159] Specifically, for the frame rate attenuation grouping, this embodiment can use the improved Newton cooling law to set the camera acquisition frame rate attenuation strategy, and the attenuation process presents an exponentially slowed attenuation, from Starts to decay, over time Decay to The frame rate change expression can be:

[0160] .

[0161] in, The sequence number in the frame rate decreasing group is The camera device in the sub-area changes with time The frame rate decay function is: is a constant coefficient. is the probability of target occurrence in this sub-region, is the second average flow of people in the sub-area in the preset past days, It is the first average flow of people in the preset past days for the frame rate decreasing group. and About time variables.

[0162] Specifically, for the frame rate increasing grouping, this embodiment can be based on an exponential gradual increase strategy, from Start to increase, after Time increases to , the specific frame rate change expression can be:

[0163] .

[0164] in, Represents the sequence number in the frame rate increasing group The function of the frame rate of the camera device in the sub-area changing with time belongs to the acquisition frame rate increase strategy. is a constant coefficient. is the sequence number of the sub-area in the frame rate increasing group, is the target occurrence probability in this sub-region. is the second average flow of people in the sub-area in the preset past days, The first average flow of people in the preset past days for the frame rate increasing group.

[0165] Optionally, when the target group is a frame rate stable group, the frame rate change expression constructed according to the target appearance probability of the sub-area includes:

[0166] Multiply the target occurrence probability of the sub-region by the first frame rate to obtain a target product, and use it as a frame rate change expression;

[0167] The frame rate of the camera device in the above instruction sub-area is changed starting from the corresponding second frame rate according to the frame rate change expression, including:

[0168] The frame rate of the camera device in the instruction sub-area is adjusted from the corresponding second frame rate and stabilized to the target product.

[0169] Specifically, the frame rate change expression can be:

[0170] .

[0171] in, The sequence number in the frame rate stable group is The frame rate of the camera device in the sub-area of ​​changes with time t, is the sequence number of the sub-region in the frame rate stable group, is the target occurrence probability in this sub-region. is the second average flow of people in the sub-area in the preset past days, The first average flow of people in the preset past days for the frame rate stability group.

[0172] Specifically, based on Figure 5 ,like Figure 6 As shown, in another target search method across camera devices proposed in this embodiment, the following steps may be included:

[0173] S601: Divide the cameras in the area into nodes to form a topology map.

[0174] S602: Calculate the distance between nodes to obtain a topological adjacency matrix, and calculate the shortest path matrix between each node.

[0175] S603: Add attributes to each node and set the master node according to the target detection result.

[0176] S604: Set different video frame acquisition strategies for different nodes.

[0177] S605: Determine the path according to the target detection result.

[0178] This embodiment proposes a camera video frame acquisition strategy based on dynamic resource allocation. The overall process is as follows: Figure 6 As shown. First, the station area is divided into multiple sub-areas, and the corresponding node topology map is constructed; then the path attribute parameters between the nodes are set according to the actual distance, and the topological adjacency matrix and the shortest path matrix between each node are calculated. Add attributes to the topological nodes corresponding to each sub-area, dynamically set the main node according to the result of target detection, set different video frame acquisition strategies and acquisition attenuation or incremental strategies for the camera devices in the sub-areas corresponding to different topological nodes, and finally confirm the path according to the result of target detection, and list the spatial order and route of the searched target in the station in chronological order, so that the staff can quickly locate the searched target.

[0179] like Figure 7 As shown, in other target search methods across camera devices proposed in this embodiment, after the process starts, this embodiment can determine whether a preset target is detected. If not, the target search can continue. If so, that is, when the preset target is searched in the frame image collected by the camera device in the sub-area corresponding to a certain topological node, the topological node is set as the main node. This embodiment can determine the distance between the main node and other surrounding topological nodes, that is, the shortest path length, that is, the shortest topological path length between the main area and other sub-areas ,judge and , The size relationship between All topological nodes are divided into node sets ,Will All topological nodes are divided into node sets ,Will All topological nodes are divided into node sets middle.

[0180] The node collection The frame rate of the camera device in the corresponding sub-area, that is, the initial video frame acquisition strategy, is set to , the node set The frame rate of the camera device in the corresponding sub-area, that is, the initial video frame acquisition strategy, is set to , the node set The frame rate of the camera device in the corresponding sub-area, that is, the initial video frame acquisition strategy, is set to .

[0181] Afterwards, this embodiment can calculate the target occurrence probability of the sub-region corresponding to each topological node in the corresponding regional group, and set the node decay strategy over time corresponding to different regional groups. , Node Collection Strategy And the node increment strategy over time . Check whether the target is found offline or moved to an area outside the station by the staff. If not, continue searching in the frame image captured by the camera device. If yes, end the search.

[0182] Specifically, this embodiment can record the time when the preset target is searched in the frame image, as well as the location information of the preset target and provide it to the staff to assist the staff in formulating offline search routes, so as to facilitate the staff to accurately search for the preset target and improve the offline search efficiency. Figure 2 For example, suppose a search target is detected at 8:00 on a certain day in the sub-area corresponding to topological node 1, and the target is detected in the sub-area corresponding to topological node 2 at 8:20, then detected in the sub-area corresponding to topological node 6 at 8:30, and detected in the sub-area corresponding to topological node 7 at 9:00. It can be concluded that the moving path of the target is to pass through the sub-areas corresponding to topological nodes 1, 2, 6 and 7 in sequence, that is, the target goes from the corridor to the hall to the fifth waiting room to the shopping gallery, and the next appearance point is likely to be in the sub-area corresponding to topological node 8. The staff can infer the approximate location of the target based on this path, thereby achieving fast and accurate search.

[0183] The cross-camera target search method proposed in this embodiment can improve the search efficiency and search quality and realize the rapid positioning of the search target. By dynamically allocating computing resources and optimizing the acquisition strategy of the camera image frame, the accuracy and success rate of the camera search are improved, and the lost personnel and items can be quickly located, which greatly improves the efficiency of the searched target and reduces the workload of the staff. The use of computing resources is optimized to avoid overloading the server system. Through the dynamic resource allocation strategy in the cross-camera search scenario, limited computing resources are reasonably used to avoid resource waste. At the same time, the target detection server is prevented from being stuck due to insufficient resources, ensuring the stable operation of the cross-camera search system. The processing efficiency of the cross-camera search system is enhanced and the response speed of the emergency time is improved. By dynamically allocating the image frame acquisition strategy of the station camera, the cross-camera search system can efficiently process a large amount of image data. The rapid positioning and tracking capabilities can record the movement trajectory of the target, so that quick action can be taken in an emergency, providing data support for post-analysis and investigation.

[0184] like Figure 8 As shown, this embodiment provides a target search system across camera devices, including:

[0185] The sending unit 101 is used to send an image acquisition instruction to the camera devices in multiple sub-areas of the target location in response to the target search instruction, so that the camera devices in each sub-area start to acquire images based on the same first frame rate to obtain the first batch of frame images; wherein the sum of each first frame rate is the preset total search resource;

[0186] A first search unit 102, configured to search for a preset target in the first batch of frame images;

[0187] A first determining unit 103 is used to determine that a preset target is searched in the first frame image, and then determine the sub-region corresponding to the first frame image as the main region;

[0188] A division unit 104, configured to divide the plurality of sub-regions into a plurality of region groups according to the spatial distance between the main region and the other sub-regions;

[0189] The second determination unit 105 is used to determine the second frame rate corresponding to the camera device in each regional group according to the spatial distance between the regional group and the main area; wherein the smaller the spatial distance between the regional group and the main area, the larger the second frame rate corresponding to the camera device in the regional group, and the sum of each second frame rate is the total search resource;

[0190] An instruction unit 106 is used to instruct the camera device in each area group to continue to collect images based on the corresponding second frame rate to obtain a second batch of frame images;

[0191] A second search unit 107, configured to search for a preset target in the second batch of frame images;

[0192] The third determining unit 108 is used to determine that if a preset target is found in the second frame image, then determine the sub-region corresponding to the second frame image as a new main region until a preset search end condition is met.

[0193] It should be noted that the processing processes of the sending unit 101, the first search unit 102, the first determination unit 103, the division unit 104, the second determination unit 105, the instruction unit 106, the second search unit 107 and the third determination unit 108 and the beneficial effects thereof can be respectively referred to in Figure 1 Steps S101 to S108 in the above are not described in detail.

[0194] Optionally, the instruction unit 106 is also used to instruct the camera devices in each area group to continue to acquire images based on the corresponding second frame rate, and divide the multiple area groups into frame rate attenuation groups, frame rate stabilization groups and frame rate increasing groups according to the spatial distance between the area groups and the main area, with the spatial distances from the main area increasing successively; respectively instruct the frame rates of the camera devices in the frame rate attenuation groups, frame rate stabilization groups and frame rate increasing groups to change from the corresponding second frame rate according to preset frame rate attenuation strategies, frame rate stabilization strategies and frame rate increasing strategies, and instruct each camera device to continue to acquire images according to the changed frame rate; wherein the sum of the frame rates of each camera device at any moment is not greater than the total search resources; and the frame images acquired by each camera device based on the second frame rate and the changed frame rate are taken as the second batch of frame images as a whole.

[0195] Optionally, the above system further includes:

[0196] The fourth determining unit is configured to:

[0197] In response to the target search instruction, before sending an image acquisition instruction to the camera devices in the multiple sub-areas of the target place, the target place is divided into multiple functional areas according to the regional function type in the target place, and each functional area is regarded as a sub-area; the location information of each sub-area, the connectivity information and the spatial distance between two sub-areas are obtained; the topological edge length between two sub-areas is calculated according to the spatial distance between two sub-areas; based on the location information of each sub-area, the connectivity information and the topological edge length between two sub-areas, a corresponding directed topological graph is constructed; based on the directed topological graph, the shortest topological path length between two sub-areas is determined;

[0198] The division unit 104 is further configured to divide the multiple sub-areas into multiple area groups according to the shortest topological path length between the main area and other sub-areas.

[0199] Optionally, the dividing unit 104 is further configured to:

[0200] Among the shortest topological path lengths between any two sub-regions, determine the maximum length; calculate a first product obtained by multiplying the maximum length by a preset first proportional coefficient, and calculate a second product obtained by multiplying the maximum length by a preset second proportional coefficient; wherein the first proportional coefficient and the second proportional coefficient are greater than 0 and less than 1, and the first proportional coefficient is less than the second proportional coefficient; among multiple sub-regions, determine a first sub-region whose shortest topological path length with the main region is less than the first product, determine a second sub-region whose shortest topological path length with the main region is greater than the first product and less than the second product, and determine a third sub-region whose shortest topological path length with the main region is greater than the second product; group all first sub-regions and the main region as a whole as a first region, group all second sub-regions as a whole as a second region, group all third sub-regions as a whole as a third region, and group the first region group, the second region group and the third region group as multiple region groups.

[0201] Optionally, the instruction unit 106 is further configured to:

[0202] The first area group, the second area group and the third area group are respectively used as a frame rate decay group, a frame rate stabilization group and a frame rate increasing group.

[0203] Optionally, the total search resources are the maximum number of frame images that can be processed per unit time when performing target search based on frame images; the first frame rate is obtained by dividing the maximum number of frame images by the total number of cameras in the target location; the second frame rate corresponding to the cameras in the first area grouping is a preset maximum allowable frame rate; the second frame rate corresponding to the cameras in the second area grouping is equal to the first frame rate; the second frame rate corresponding to the cameras in the third area grouping is a preset minimum allowable frame rate; wherein the first frame rate is greater than the minimum allowable frame rate and less than the maximum allowable frame rate.

[0204] Optionally, the instruction unit 106 is further configured to:

[0205] Determine the first average flow of people in the target group in the preset past number of days, the target group is a frame rate attenuation group, a frame rate stable group or a frame rate increasing group; for any sub-area in the target group, the second average flow of people in the sub-area in the preset past number of days, divide the second average flow of people by the first average flow of people to obtain the target appearance probability of the sub-area, construct a frame rate change expression according to the target appearance probability of the sub-area, and instruct the frame rate of the camera device in the sub-area to change according to the frame rate change expression starting from the corresponding second frame rate.

[0206] Optionally, when the target grouping is a frame rate attenuation grouping, a frame rate change expression constructed based on the target occurrence probability of the sub-area in the target grouping is used to attenuate the frame rate of the camera device in the sub-area from the corresponding second frame rate to the first frame rate; when the target grouping is a frame rate increasing grouping, a frame rate change expression constructed based on the target occurrence probability of the sub-area in the target grouping is used to increase the frame rate of the camera device in the sub-area from the corresponding second frame rate to the first frame rate.

[0207] Optionally, when the target group is a frame rate stable group, the instruction unit 106 is further configured to: multiply the target appearance probability of the sub-region by the first frame rate to obtain a target product, and use the product as a frame rate change expression;

[0208] The instruction unit 106 is further used to instruct the frame rate of the camera device in the sub-area to be adjusted from the corresponding second frame rate and stabilized to the target product.

[0209] The cross-camera target search system proposed in this embodiment can set reasonable image acquisition frame rates for cameras in different areas of the target location, perform target search on frame images acquired by the cameras based on the corresponding frame rates, and dynamically adjust the frame rates of cameras in different areas based on real-time target search results. In a cross-camera search scenario, dynamic allocation and optimization of search resources can be performed, and rapid search for preset targets can be achieved while reasonably allocating search resources and fully utilizing them, thereby reducing the human resources consumed by manual search and improving the search efficiency for preset targets.

[0210] The target search system across camera devices in this embodiment is presented in the form of functional units, where the units refer to ASIC (Application Specific Integrated Circuit) circuits, processors and memories that execute one or more software or fixed programs, and / or other devices that can provide the above functions. The embodiment of the present invention also provides a computer device having the above Figure 8 The target search system across the camera device is shown.

[0211] See also Fig. 9, a schematic diagram of the structure of a computer device provided by an optional embodiment of the present invention, the computer device includes: one or more processors 10, a memory 20, and interfaces for connecting various components, including high-speed interfaces and low-speed interfaces. The various components are connected to each other using different buses for communication, and can be installed on a common motherboard or installed in other ways as needed. The processor can process instructions executed in the computer device, including instructions stored in or on the memory to display graphical information of the GUI on an external input / output device (such as a display device coupled to the interface). In some optional embodiments, if necessary, multiple processors and / or multiple buses can be used together with multiple memories and multiple memories. Similarly, multiple computer devices can be connected, and each device provides part of the necessary operations (for example, as a server array, a group of blade servers, or a multi-processor system). Fig. 9 A processor 10 is taken as an example.

[0212] The processor 10 may be a central processing unit, a network processor or a combination thereof. The processor 10 may further include a hardware chip. The hardware chip may be a dedicated integrated circuit, a programmable logic device or a combination thereof. The programmable logic device may be a complex programmable logic device, a field programmable gate array, a general array logic or any combination thereof. The memory 20 stores instructions executable by at least one processor 10 so that at least one processor 10 executes the method shown in the above embodiment.

[0213] The memory 20 may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function. The data storage area may store data created according to the use of the computer device, etc. In addition, the memory 20 may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage devices. In some optional embodiments, the memory 20 may optionally include a memory remotely arranged relative to the processor 10, and these remote memories may be connected to the computer device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and a combination thereof. The memory 20 may include a volatile memory, such as a random access memory. The memory may also include a non-volatile memory, such as a flash memory, a hard disk, or a solid-state hard disk. The memory 20 may also include a combination of the above-mentioned types of memories.

[0214] The computer device further comprises a communication interface 30 for the computer device to communicate with other devices or a communication network.

[0215] The embodiment of the present invention also provides a computer-readable storage medium. The method according to the embodiment of the present invention can be implemented in hardware, firmware, or can be implemented as a computer code that can be recorded in a storage medium, or can be implemented as a computer code that is originally stored in a remote storage medium or a non-temporary machine-readable storage medium and will be stored in a local storage medium through a network download, so that the method described herein can be stored in such software processing on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only storage memory, a random access memory, a flash memory, a hard disk or a solid-state hard disk, etc.; further, the storage medium can also include a combination of the above types of memories. It can be understood that a computer, a processor, a microprocessor controller, or programmable hardware includes a storage component that can store or receive software or computer code. When the software or computer code is accessed and executed by a computer, a processor, or hardware, the method shown in the above embodiment is implemented.

[0216] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A target search method across camera devices, characterized in that: include: In response to the target search instruction, an image acquisition instruction is sent to the camera devices in the multiple sub-areas of the target location, so that the camera devices in each of the sub-areas start to acquire images based on the same first frame rate to obtain the first batch of frame images; wherein the sum of each of the first frame rates is the preset total search resource; Searching for a preset target in the first batch of frame images, and determining that the preset target is found in the first frame image, and determining the sub-region corresponding to the first frame image as the main region; Dividing the plurality of sub-regions into a plurality of region groups according to the spatial distance between the main region and the other sub-regions; Determine, according to the spatial distance between the area group and the main area, a second frame rate corresponding to the camera device in each of the area groups; wherein, the smaller the spatial distance between the area group and the main area, the larger the second frame rate corresponding to the camera device in the area group, and the sum of each of the second frame rates is the total search resource; Instructing the camera device in each of the area groups to continue to collect images based on the corresponding second frame rate to obtain a second batch of frame images; Searching for the preset target in the second batch of frame images, and determining that the preset target is found in the second frame image, determining the sub-region corresponding to the second frame image as a new main region, until a preset search end condition is met; The instructing the camera device in each of the area groups to continue to acquire images based on the corresponding second frame rate to obtain a second batch of frame images includes: Instructing the camera device in each of the area groups to continue to acquire images based on the corresponding second frame rate, and dividing the plurality of area groups into a frame rate decaying group, a frame rate stabilizing group, and a frame rate increasing group, the spatial distances between the area groups and the main area increasing in sequence, according to the spatial distances between the area groups and the main area; Respectively instruct the frame rates of the camera devices in the frame rate decay group, the frame rate stabilization group, and the frame rate increment group to change according to the preset frame rate decay strategy, frame rate stabilization strategy, and frame rate increment strategy starting from the corresponding second frame rate, and instruct each camera device to continue to collect images according to the changed frame rate; wherein the sum of the frame rates of each camera device at any time is not greater than the total search resource; The frame images collected by each camera device based on the second frame rate and the changed frame rate are taken as the second batch of frame images.

2. The method according to claim 1, characterized in that: Before sending an image acquisition instruction to the camera devices in the plurality of sub-areas of the target location in response to the target search instruction, the method further includes: According to the regional functional type in the target place, the target place is divided into a plurality of functional areas, and each functional area is used as a sub-area; Obtaining location information of each sub-region, connectivity information between two sub-regions, and spatial distance; Calculating the topological side lengths between the sub-regions according to the spatial distances between the sub-regions; Based on the location information of each sub-region, the connectivity information between two sub-regions and the topological edge length, a corresponding directed topological graph is constructed; Based on the directed topological graph, determining the shortest topological path length between each of the sub-areas; The dividing the plurality of sub-regions into a plurality of region groups according to the spatial distance between the main region and the other sub-regions comprises: The multiple sub-areas are divided into multiple area groups according to the shortest topological path length between the main area and other sub-areas.

3. The method according to claim 2, characterized in that The dividing the plurality of sub-areas into a plurality of area groups according to the topological path lengths between the main area and the other sub-areas comprises: Among the shortest topological path lengths between the sub-areas, a maximum length is determined; Calculating a first product obtained by multiplying the maximum length by a preset first proportional coefficient, and calculating a second product obtained by multiplying the maximum length by a preset second proportional coefficient; wherein the first proportional coefficient and the second proportional coefficient are greater than 0 and less than 1, and the first proportional coefficient is less than the second proportional coefficient; Among the multiple sub-areas, determine a first sub-area whose shortest topological path length with the main area is less than the first product, determine a second sub-area whose shortest topological path length with the main area is greater than the first product and less than the second product, and determine a third sub-area whose shortest topological path length with the main area is greater than the second product; All the first sub-regions and the main region as a whole are grouped as a first region, all the second sub-regions as a whole are grouped as a second region, all the third sub-regions as a whole are grouped as a third region, and the first region group, the second region group and the third region group are grouped as the plurality of region groups.

4. The method according to claim 3, characterized in that The method of dividing the plurality of regional groups into a frame rate attenuation group, a frame rate stabilization group, and a frame rate increasing group, whose spatial distances from the main region increase in sequence, according to the spatial distances between the regional groups and the main region, comprises: The first area group, the second area group and the third area group are respectively used as the frame rate decay group, the frame rate stabilization group and the frame rate increasing group.

5. The method according to claim 3 or 4, characterized in that: The total search resource is the maximum number of frame images that can be processed within a unit time when performing target search based on frame images; The first frame rate is obtained by dividing the maximum number of image frames by the total number of camera devices in the target location; The second frame rate corresponding to the camera device in the first area group is a preset maximum allowable frame rate; A second frame rate corresponding to the camera device in the second area group is equal to the first frame rate; The second frame rate corresponding to the camera device in the third area group is a preset minimum allowed frame rate; The first frame rate is greater than the minimum allowed frame rate and less than the maximum allowed frame rate.

6. The method according to claim 5, characterized in that The respectively instructing the frame rates of the camera devices in the frame rate decay group, the frame rate stabilization group, and the frame rate increasing group to change according to a preset frame rate decay strategy, frame rate stabilization strategy, and frame rate increasing strategy starting from the corresponding second frame rate, comprises: Determine a first average flow of people of a target group in a preset number of past days, wherein the target group is the frame rate decaying group, the frame rate stable group, or the frame rate increasing group; For any of the sub-areas in the target group, the second average flow of people in the sub-area in the preset past days is obtained by dividing the second average flow of people by the first average flow of people to obtain the target appearance probability of the sub-area, and a frame rate change expression is constructed according to the target appearance probability of the sub-area. The frame rate of the camera device in the sub-area is instructed to change according to the frame rate change expression starting from the corresponding second frame rate.

7. The method according to claim 6, characterized in that When the target group is the frame rate attenuation group, a frame rate change expression constructed according to the target appearance probability of the sub-area in the target group is used to attenuate the frame rate of the camera device in the sub-area from the corresponding second frame rate to the first frame rate; When the target group is the frame rate increasing group, the frame rate change expression constructed according to the target appearance probability of the sub-area in the target group is used to increase the frame rate of the camera device in the sub-area from the corresponding second frame rate to the first frame rate.

8. The method according to claim 6, characterized in that When the target group is the frame rate stable group, constructing a frame rate change expression according to the target appearance probability of the sub-area includes: Multiplying the target occurrence probability of the sub-region by the first frame rate to obtain a target product, and using the target product as the frame rate change expression; The instructing the frame rate of the camera device in the sub-area to change from the corresponding second frame rate according to the frame rate change expression includes: The frame rate of the camera in the sub-area is instructed to be adjusted from the corresponding second frame rate and stabilized to the target product.

9. A target search system across camera devices, characterized in that: include: A sending unit, configured to send an image acquisition instruction to the camera devices in the plurality of sub-areas of the target location in response to the target search instruction, so that the camera devices in each of the sub-areas start to acquire images based on the same first frame rate to obtain the first batch of frame images; wherein the sum of each of the first frame rates is a preset total search resource; A first search unit, used for searching for a preset target in the first batch of frame images; A first determining unit, configured to determine that the preset target is searched in the first frame image, and then determine the sub-region corresponding to the first frame image as a main region; a division unit, configured to divide the plurality of sub-regions into a plurality of region groups according to the spatial distance between the main region and the other sub-regions; A second determination unit is used to determine a second frame rate corresponding to the camera device in each of the area groups according to a spatial distance between the area group and the main area; wherein the smaller the spatial distance between the area group and the main area, the larger the second frame rate corresponding to the camera device in the area group, and the sum of each of the second frame rates is the total search resource; An instruction unit, used for instructing the camera device in each of the area groups to continue to collect images based on the corresponding second frame rate to obtain a second batch of frame images; A second search unit, configured to search for the preset target in the second batch of frame images; a third determining unit, configured to determine that if the preset target is found in the second frame image, the sub-region corresponding to the second frame image is determined as a new main region until a preset search end condition is met; Wherein, the instruction unit is also used to: instruct the camera device in each of the area groups to continue to acquire images based on the corresponding second frame rate, and divide the multiple area groups into frame rate attenuation groups, frame rate stabilization groups and frame rate increasing groups with increasing spatial distances from the main area according to the spatial distance between the area groups and the main area; respectively instruct the frame rates of the camera devices in the frame rate attenuation groups, the frame rate stabilization groups and the frame rate increasing groups to change according to the preset frame rate attenuation strategy, frame rate stabilization strategy and frame rate increasing strategy starting from the corresponding second frame rate, and instruct each camera device to continue to acquire images according to the changed frame rate; wherein the sum of the frame rates of each camera device at any time is not greater than the total search resources; and the frame images acquired by each camera device based on the second frame rate and the changed frame rate are taken as the second batch of frame images as a whole.

Images