Method and device for estimating object distribution information, electronic equipment and storage medium

Abstract

The present disclosure provides a method and device for estimating object distribution information, electronic equipment and storage medium, relating to the technical field of artificial intelligence, in particular to the fields of intelligent transportation, computer vision, image processing, etc. The specific implementation scheme is: for each image in a plurality of images, determining object information of an object in the image, and determining first position information of the object in the image; the plurality of images are obtained by an image acquisition device collecting a region; according to the first position information, determining second position information of the object in map data; and according to the second position information, determining estimated object distribution information of a plurality of sub-regions in the region at a target time after a current time.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to the technical field of artificial intelligence, in particular to the fields of intelligent transportation, computer vision, image processing, and the like, and more specifically, the present disclosure provides a method and apparatus for estimating object distribution information, an electronic device, a storage medium, and a computer program product. BACKGROUND

[0002] People gradually take camping in an open environment as a new way of travel. It is difficult for the managers of the camping sites to manage and count the personnel flow in the camping sites, and consumers cannot master the passenger flow of the destination and cannot obtain accurate camping sites, thereby increasing the management difficulty and affecting the travel experience. SUMMARY

[0003] The present disclosure provides a method and apparatus for estimating object distribution information, an electronic device, a storage medium, and a computer program product.

[0004] According to an aspect of the present disclosure, a method for estimating object distribution information is provided, including: determining, for each image in a plurality of images, object information of an object in the image and first position information of the object in the image; the plurality of images are obtained by an image acquisition device collecting a region; determining, according to the first position information, second position information of the object in map data; and determining, according to the second position information, estimated object distribution information of a plurality of sub-regions in the region at a target time after a current time.

[0005] According to another aspect of the present disclosure, an apparatus for estimating object distribution information is provided, including: an information determining module, a second position determining module, and a distribution information determining module. The information determining module is configured to determine, for each image in a plurality of images, object information of an object in the image and first position information of the object in the image; the plurality of images are obtained by an image acquisition device collecting a region; the second position determining module is configured to determine, according to the first position information, second position information of the object in map data; and the distribution information determining module is configured to determine, according to the second position information, estimated object distribution information of a plurality of sub-regions in the region at a target time after a current time.

[0006] According to another aspect of the present disclosure, an electronic device is provided, including: at least one processor; and a memory communicatively connected with the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method provided by the present disclosure.

[0007] According to another aspect of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, where the computer instructions are used to cause a computer to execute the method provided by the present disclosure.

[0008] According to another aspect of the present disclosure, a computer program product is provided, comprising a computer program which, when executed by a processor, implements the method provided by the present disclosure.

[0009] It should be understood that the contents described in this section are not intended to identify key or important features of the embodiments of the present disclosure, nor to limit the scope of the present disclosure. Other features of the present disclosure will become apparent through the following description. BRIEF DESCRIPTION OF DRAWINGS

[0010] The accompanying drawings are used to better understand the present scheme, and do not limit the present disclosure. Among them:

[0011] Figure 1 is an application scenario diagram of the method and device for estimating object distribution information according to the embodiments of the present disclosure;

[0012] Figure 2 is a schematic flowchart of the method for estimating object distribution information according to the embodiments of the present disclosure;

[0013] Figure 3 is a schematic principle diagram of the processing process for the first object according to the embodiments of the present disclosure;

[0014] Figure 4 is a schematic flowchart of the method for determining the estimated object distribution information according to the second position information according to the embodiments of the present disclosure;

[0015] Figure 5 is a schematic structural diagram of a point-edge graph according to the embodiments of the present disclosure;

[0016] Figure 6 is a schematic diagram of an object change curve according to the embodiments of the present disclosure;

[0017] Figure 7 is a schematic structural diagram of a system for estimating object distribution information according to the embodiments of the present disclosure;

[0018] Figure 8 is a schematic structural block diagram of a device for estimating object distribution information according to the embodiments of the present disclosure; and

[0019] Figure 9 is a structural block diagram of an electronic device for implementing the method for estimating object distribution information according to the embodiments of the present disclosure. DETAILED DESCRIPTION

[0020] Exemplary embodiments of the present disclosure are described herein with reference to the accompanying drawings, which are meant to be exemplary. Therefore, various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Also, descriptions of known functions and constructions are omitted for clarity and conciseness.

[0021] In some embodiments, the management personnel can realize the management and statistics of personnel flow through the sale of tickets, the entry and exit records of vehicles, manual inspection, etc., but the utilization of various areas of the campsite is unbalanced, and the use of the campsite cannot be accurately counted and managed. Consumers cannot obtain information about the campsite, lack of open space guidance, and spend a lot of time when looking for open space, which reduces the travel experience.

[0022] It can be seen that the above embodiments obtain information flow from the whole, and for the camping scene, local information is more important. The acquisition and display of local information not only facilitates the planning of the scenic area, but also enables consumers to find the best camping site more quickly and effectively.

[0023] The embodiments of the present disclosure aim to provide a method for estimating object distribution information, which can use an image acquisition device to collect images in real time, determine the tent position, number of personnel and other information of each campsite based on the images, and analyze the camping situation in the area, thereby providing comprehensive local campsite information for the management personnel and consumers.

[0024] The technical solutions provided by the present disclosure will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0025] Figure 1 is a schematic diagram of an application scenario of the method and device for estimating object distribution information according to the embodiments of the present disclosure.

[0026] It should be noted that Figure 1 The system architecture shown is only an example of a system architecture to which the embodiments of the present disclosure can be applied, to help those skilled in the art understand the technical content of the present disclosure, but it does not mean that the embodiments of the present disclosure cannot be used in other devices, systems, environments or scenarios.

[0027] As Figure 1 shown, the system architecture 100 according to the embodiments can include terminal devices 101, 102, 103, a network 104 and a server 105. The network 104 is used as a medium to provide a communication link between the terminal devices 101, 102, 103 and the server 105. The network 104 can include various connection types, such as wired and / or wireless communication links, etc.

[0028] The user can use the terminal devices 101, 102, 103 to interact with the server 105 through the network 104 to receive or send messages, etc. The terminal devices 101, 102, 103 can be various electronic devices with display screens and supporting web browsing, including but not limited to smart phones, tablet computers, laptop portable computers, desktop computers, etc.

[0029] The server 105 can be a server providing various services, for example, a background management server supporting a website browsed by the user using the terminal devices 101, 102, 103 (only as an example). The background management server can analyze and process the received user request data, etc., and feed back the processing result (for example, the estimated object distribution information such as the crowd flow, the pedestrian density, the tent density, etc. of each sub-region in the region obtained or generated according to the user request) to the terminal device.

[0030] It should be noted that the method for estimating the object distribution information provided by the embodiments of the present disclosure can generally be executed by the server 105. Accordingly, the apparatus for estimating the object distribution information provided by the embodiments of the present disclosure can generally be arranged in the server 105. The method for estimating the object distribution information provided by the embodiments of the present disclosure can also be executed by a server or a server cluster different from the server 105 and capable of communicating with the terminal devices 101, 102, 103 and / or the server 105. Accordingly, the apparatus for estimating the object distribution information provided by the embodiments of the present disclosure can also be arranged in a server or a server cluster different from the server 105 and capable of communicating with the terminal devices 101, 102, 103 and / or the server 105.

[0031] It should be understood that Figure 1 The number of terminal devices, networks and servers in

[0032] Figure 2 is a schematic flowchart of the method for estimating the object distribution information according to the embodiments of the present disclosure.

[0033] As shown in Figure 2 , the method 200 for estimating the object distribution information can include operation S210 to operation S230.

[0034] In operation S210, for each of the plurality of images, object information of an object in the image is determined, and first position information of the object in the image is determined; the plurality of images are obtained by an image acquisition device collecting the region.

[0035] For example, the region can be a park, a scenic spot, or the like. The image acquisition device can be a camera, such as a high-position camera. The number of image acquisition devices can be one or multiple. When the number of image acquisition devices is multiple, the collection fields of view of the multiple image acquisition devices can or can not have an intersection.

[0036] For example, the object can include a moving object and a stationary object. The moving object can include a pedestrian, a pet, a vehicle (such as a scooter, a car, or the like), or the like. The stationary object can include a tent or the like.

[0037] For example, target detection can be performed on the image to obtain a detection box of the object. Based on the detection box, the object information and the first position information of the object can be determined. For example, the object information of the object can include the category and the identification of the object. The first position information of the object in the image can be embodied by the length, the height, and the center point coordinate of the detection box.

[0038] In operation S220, the second position information of the object in the map data is determined according to the first position information.

[0039] For example, since the installation position and the collection field of view of the image acquisition device are fixed, the image acquisition device can be calibrated in advance to determine the physical position in the map data corresponding to each pixel in the image collected by the image acquisition device, so as to obtain the mapping relationship between the first position information and the second position information. Then, the first position information is converted into the second position information based on the mapping relationship.

[0040] In operation S230, the estimated object distribution information of each of the multiple sub-regions in the region at a target time after the current time is determined according to the second position information.

[0041] For example, the object distribution information can include the total amount of objects and the object density. For example, the estimated object distribution information can include the number of pedestrians, the density of pedestrians, the number of tents, the density of tents, or the like. The object distribution information at the current time is referred to as initial object distribution information, and the object distribution information at the target time is referred to as estimated object distribution information. The target time can be any time after the current time. The time interval between the target time and the current time can be small.

[0042] For example, the region can be divided into multiple sub-regions in advance. Then, the initial object distribution information of each sub-region at the current time can be counted. The initial object distribution information can include the total amount of objects and the object density in each sub-region. The initial object distribution information at the current time can be used as the estimated object distribution information at the target time, so as to reduce the data processing amount and improve the processing efficiency.

[0043] For example, a point-edge graph can be generated according to the second position information, and then the estimated object distribution information can be determined based on the point-edge graph. Compared with a manner of taking the initial object distribution information at the current time as the estimated object distribution information at the target time, the manner of determining the estimated object distribution information based on the point-edge graph can improve the accuracy. The process of generating the point-edge graph and determining the estimated object distribution information based on the point-edge graph will be described in detail below, and thus will not be described here again.

[0044] In actual applications, after obtaining the estimated object distribution information of each sub-region, the estimated object distribution information of each sub-region can be output to a terminal device, such as a mobile phone, used by a user, and visualized.

[0045] The embodiments of the present disclosure use an image acquisition device to acquire an image of a region, then analyze the image to determine first position information of an object in the image, subsequently convert the first position information into second position information in map data, and determine estimated object distribution information of each sub-region in the future based on the second position information, thereby facilitating the user to know the information such as the flow and density of people in the region, and providing the user and the management party with guidance services for idle sub-regions, so that the user can plan his / her own travel according to the actual situation and relieve the pressure of people flow.

[0046] Figure 3 is a schematic principle diagram of a processing process for a first object according to the embodiments of the present disclosure.

[0047] According to another embodiment of the present disclosure, the object can include a first object, and the first object can be a mobile object, such as a pedestrian, a pet, a scooter, a car, etc. The process of determining the object information of the first object will be described below.

[0048] It should be noted that taking the mobile object as a pedestrian as an example, the personnel are flowing within a certain range, and a single image acquisition device cannot match the multi-scene trajectory of the same person. Therefore, multiple image acquisition devices can be used to acquire images of the entire region. When the collection fields of view of the multiple image acquisition devices have intersections, different image acquisition devices can capture the first object at the same position. Therefore, in order to ensure the accuracy of the object, a deduplication process needs to be performed.

[0049] If Figure 3 As shown in FIG. 3, target detection can be performed on the image 301 to obtain a detection frame 302 for the first object in the image 301, and then the feature 303 of the first object is determined according to the detection frame 302. Subsequently, the similarity 305 between the feature 303 of the first object and the features 304 of multiple candidate first objects is determined, and the identification 306 of the first object is determined according to the similarity 305, and the identification 306 of the first object is taken as the object information of the first object.

[0050] For example, a deep learning-based convolutional neural network (CNN) can be used to extract the features 303 of the first object.

[0051] For another example, a feature extraction network including a Transformer can be used to input the pedestrian image 301 into the network and extract the features 303 of the first object in the image 301. The features 303 of the first object can be local features (e.g., a face) of the image 301, but in the case of a person facing away from the camera, there may be a mismatch or even no match, so the features 303 of the first object can be global features, thereby achieving 360-degree matching and improving the accuracy of matching.

[0052] It should be noted that the cross-camera pedestrian re-identification based on the Transformer is a technology for identifying pedestrians under different image capture devices. The Transformer-based method utilizes the self-attention mechanism of the Transformer network, which can model the input sequence globally and capture the dependency between elements in the sequence. In cross-camera pedestrian re-identification, this capability is very useful for learning and matching pedestrian features, as pedestrians may have different perspectives, lighting conditions, and scales under different cameras. During the training of the Transformer network, pedestrian image data with different perspectives can be collected and preprocessed, such as cropping, scaling, and normalization. AIGC can be used to generate images of the same image 301 under multiple perspectives, and then the Transformer network is trained using these images.

[0053] For example, for a given image 301 of a first object, the features 303 of the first object are compared with the features 304 of multiple candidate first objects in the database, and the similarity 305 is calculated. The similarity 305 can be calculated using a metric learning method (e.g., cosine similarity, Euclidean distance) or a contrastive loss-based method.

[0054] For example, after calculating the similarity 305, the candidate first objects can be sorted according to the calculated similarity 305. If the maximum similarity is greater than or equal to a first similarity threshold, the identification of the candidate first object corresponding to the maximum similarity is determined as the identification 306 of the first object. If all similarities are less than the first similarity threshold, it indicates that the database lacks the first object, and other identifications other than the identifications of the candidate objects can be generated as the identification 306 of the first object. The first similarity threshold can be 0.8, and the present embodiment does not limit the first similarity threshold. In this way, the identification 306 of the first object can be accurately determined.

[0055] After obtaining the identification 306 of the first object, the identification 306 of the first object at the same time is counted, and the number 307 of the first object can be obtained. For the same first object, the second position information of the first object at multiple times is fitted, and the trajectory 308 of the first object can be obtained.

[0056] The embodiment uses multiple image collection devices to collect images 301 of the region, and the identification 306 of the first object is de-duplicated according to the similarity 305, so as to ensure accurate counting of the number of the first object and accurate determination of the trajectory of the first object.

[0057] According to another embodiment of the present disclosure, the object can include a second object, and the second object can be a stationary object such as a tent, and the following describes a process of determining object information of the second object.

[0058] It should be noted that taking the tent as an example of the moving object, although the mobility of the tent is lower than that of the pedestrian, there is still a problem that a single image collection device cannot collect all the tents in the region, and therefore, multiple image collection devices can be used to collect images of the entire region. When the collection fields of view of the multiple image collection devices have intersections, different image collection devices can capture the second object at the same position. Due to algorithm errors and other factors, the second position information of the second object obtained by processing different images can be deviated, and the same second object can be counted multiple times. Therefore, in order to ensure the accuracy of the object, de-duplication processing can be performed.

[0059] For example, in a case where it is detected that two second objects satisfy a predetermined condition, the two second objects can be determined as the same object, at this time, one of the two second objects can be deleted to realize de-duplication, and an identification of another second object of the two second objects is generated, and the identification of the second object is taken as the object information of the second object.

[0060] For example, the predetermined condition can include that the distance between the two second objects in the map data is less than or equal to a predetermined distance. In addition, the predetermined condition can also include that the similarity between two features of the two second objects is greater than or equal to a second similarity threshold, and the second similarity threshold can be 0.8, and the embodiment does not limit the second similarity threshold.

[0061] The embodiment uses multiple image collection devices to collect images of the region, and de-duplicates the second object according to the predetermined condition, so as to accurately count the number of the second object.

[0062] Figure 4 is a schematic flowchart of a method for determining estimated object distribution information according to second position information according to an embodiment of the present disclosure.

[0063] In this embodiment, the method 430 of determining the estimated object distribution information of each sub-region in the region at the target time after the current time according to the second position information can include operation S431 to operation S434.

[0064] In operation S431, the second position information of the plurality of objects in the plurality of images in the map data is clustered to obtain a plurality of clusters, each cluster corresponding to a sub-region in the region.

[0065] For example, the second position information of each object can be mapped to the map data, and then clustered to divide the positions of the plurality of objects into a plurality of discrete sub-regions, at least one second position information close in distance forming a cluster, so as to obtain a plurality of clusters. The embodiment does not limit the clustering manner. Each cluster includes at least one second position information, and the cluster corresponds to a sub-region including all second position information in the cluster.

[0066] In other embodiments, the region can also be divided into a plurality of discrete sub-regions by average division or according to a preconfigured division manner.

[0067] In operation S432, a plurality of nodes corresponding to the plurality of clusters are generated in the point-edge graph.

[0068] For example, the point-edge graph includes nodes, and a node is generated for each cluster.

[0069] For example, the node can have attribute information, for example, the attribute information can include cluster center position information, initial object density, and historical change information, etc. The cluster center position information represents the position information of the sub-region corresponding to the node in the map data. The initial object density represents the object density of the sub-region corresponding to the node at the current time, and the initial object density can be updated every predetermined time length, which can be 15 minutes. The historical change information represents the object density change of the sub-region corresponding to the node, which can be density change amount, density change rate, etc.

[0070] In operation S433, an edge in the point-edge graph is generated according to the map data and the plurality of sub-regions corresponding to the plurality of clusters.

[0071] For example, the point-edge graph further includes edges, two nodes are connected via an edge, the edge can be an undirected edge or a directed edge. When the edge is a directed edge, the direction of the edge can represent the direction of object flow between the nodes. For example, there are objects moving from the sub-region corresponding to the first node to the sub-region corresponding to the second node, and the edge can be directed from the first node to the second node, and the direction of the edge can be bidirectional. In addition, the direction of the edge can be unidirectional at the same time, and the direction of the edge is consistent with the direction of more trajectories, for example, there are 100 trajectories moving from the first node to the second node, and there are 20 trajectories moving from the second node to the first node, and the unidirectional edge can be directed from the first node to the second node.

[0072] In operation S434, according to the point-edge graph, the estimated object distribution information of each of the plurality of sub-regions is determined.

[0073] Density estimation can be performed based on the graph, for example, using a graph data structure and a density algorithm to estimate the crowd density, and the estimation process is based on the following assumptions: the crowd density has correlation and locality in space, and the area with high crowd density is often surrounded by high crowd density.

[0074] The embodiment determines the estimated object distribution information based on the point-edge graph, thereby improving the accuracy of the estimated object distribution information.

[0075] Figure 5 is a schematic structural diagram of a point-edge graph according to an embodiment of the present disclosure.

[0076] As shown in Figure 5 In the embodiment, the point-edge graph includes nodes A, B, C, and D, each node corresponds to a sub-region, node A is connected to nodes B, C, and D via an edge, indicating that the sub-region of node A is close to the sub-regions of nodes B, C, and D or on the same path. Node C is connected to nodes B and D via an edge, indicating that the sub-region of node C is close to the sub-regions of nodes B and D or on the same path. The direction of the directed edge indicates that more objects move from one node to another node, for example, more objects move from node C to nodes A, B, and D.

[0077] According to another embodiment of the present disclosure, the edges in the point-edge graph can be generated in various ways, and the process of generating the edges is described below.

[0078] In one example, for any two nodes in the plurality of nodes, a link can be generated between the two nodes in response to detecting that the two cluster center position information of the two nodes are in a same path in the map data, and the two cluster center position information are adjacent in the same path. For example, the two cluster center position information being in a same path in the map data indicates that the first sub-region can reach the second sub-region through the path. Adjacent indicates that the two sub-regions are adjacent to each other in the same path. This embodiment generates the link according to the spatial distance between the sub-regions, so that the point-link graph is consistent with the physical spatial relationship between the sub-regions.

[0079] In another example, for any two nodes in the plurality of nodes, a link can be generated between the two nodes in response to detecting that the path distance of the two cluster center position information of the two nodes in the map data is less than or equal to a distance threshold. For example, the path distance of the two cluster center position information in the map data being less than or equal to the distance threshold indicates that the path distance from one sub-region to another sub-region is relatively short. This embodiment generates the link according to the spatial distance between the sub-regions, so that the point-link graph is consistent with the physical spatial relationship between the sub-regions.

[0080] In another example, for any two nodes in the plurality of nodes, a link can be generated between the two nodes in response to detecting that the number of object trajectories passing through the two nodes is greater than or equal to a number threshold.

[0081] For example, the object trajectory is determined according to the second position information of the object at a plurality of time points. For example, the second position information of a single object at a plurality of time points is fitted to obtain the trajectory of the object, and the object can be the first object.

[0082] For example, the point-link graph includes a first node and a second node, in which there are 100 trajectories moving from the first node to the second node, and there are 20 trajectories moving from the second node to the first node. Since the number of trajectories connecting the first node and the second node is 120, which is relatively large, a link can be generated between the first node and the second node. In addition, the direction of the link can be that the first node points to the second node. The link is generated by analyzing the trajectories, so that the point-link graph is consistent with the movement trajectories of the user.

[0083] The above describes the process of generating the link. It should be noted that in actual application, any one of the above embodiments can be used to generate the link, or a plurality of the above embodiments can be combined to generate the link.

[0084] According to another embodiment of the present disclosure, the method for determining the estimated object distribution information of each of the plurality of sub-regions according to the point-edge graph can include the following process: for each node in the point-edge graph, determining a density change amount of the node according to the historical change information of the node, and then determining the estimated object density of the sub-region corresponding to the node at the target time according to the initial object density of the node and the density change amount.

[0085] For example, the initial object density of the sub-region corresponding to each node can be calculated according to the crowd data at the current time, and the initial object density can be measured by the number of people, flow or other indicators measuring the crowd, for example, the ratio of the area of the sub-region to the number of people can be taken as the initial object density.

[0086] The following describes the process of determining the density change amount of the first node, which is any one of the plurality of nodes, and the first node is connected to the second node via an edge

[0087] In an embodiment, the mapping relationship between the ratio and the density change amount can be pre-configured, if the ratio is greater than 1, the density change rate is positive; if the ratio is less than 1, the density change rate is negative; if the ratio is less than 1, the density change rate is 0, in addition, the density change rate can be positively correlated with the ratio.

[0088] The number of first trajectories from the first node to the second node can be determined, and the number of second trajectories from the second node to the first node can be determined, then the ratio between the number of first trajectories and the number of second trajectories can be determined, and the density change amount corresponding to the ratio can be found based on the mapping relationship. If the first node is connected to a plurality of second nodes via an edge, the sum of the density change amounts obtained based on all the second nodes can be taken as the density change amount for the first node, that is, the total amount of inflow and outflow is used to count the density change amount of the first node.

[0089] In another example, according to the cross-lens recognition result, the direction of the crowd between the nodes can be determined, and the flow information between the two nodes is counted in the time dimension, for example, a curve is fitted with time as the horizontal coordinate and the object change amount as the vertical coordinate, as shown in Figure 6 .

[0090] In practical applications, a pre-trained model can be used to determine the object change amount, for example, the input of the model can include the historical change amount of the past 1 hour, which is counted at a frequency of every minute, then the model is a vector with an input of 1*60. The output of the model is the object density change amount at the next time. The model can be, for example, an xgboost tree model. For another example, a time series analysis algorithm based on RCNN can be used to estimate the crowd change at the next time.

[0091] For example, the sum of the initial object density and the density change amount can be determined as the estimated object density.

[0092] The embodiment propagates the density information in the point-edge graph using the density propagation method based on the edge and node relationship in the point-edge graph. The propagation process can update the object density of the sub-region corresponding to each node through the interaction between the nodes and the influence of the adjacent nodes, thereby improving the accuracy of estimating the object distribution information.

[0093] Figure 7 FIG. 1 is a schematic structural diagram of a system for estimating object distribution information according to an embodiment of the present disclosure.

[0094] In the embodiment, computer vision and artificial intelligence technologies are used to estimate the density of objects such as pedestrians and tents. The embodiment relates to a data source system 750, a data processing system 740, an information processing system 730, a data platform 720, and an information presentation system 710.

[0095] The data source system 750 can obtain a video data stream collected by a high-position camera, and the video data stream includes multiple images. Map data can also be obtained.

[0096] The data processing system 740 is configured to transmit and store the obtained video data stream.

[0097] The information processing system 730 can use computer vision and artificial intelligence technologies to analyze the personnel in the scene and locate the tents. Personnel analysis can include pedestrian quantity statistics, trajectory analysis, and other content. Personnel analysis can use deep learning-based human body detection and tracking technology, and tent positioning can use deep learning-based target detection technology and monocular vision-based distance estimation technology. The personnel analysis and tent positioning method can refer to the above, and the embodiment will not be described again.

[0098] The data platform 720 can process the obtained personnel information and location information, map the first location information to the second location information in the map data, and generate a data chart based on the obtained data for data display.

[0099] The information presentation system 710 can distribute the obtained personnel information and campsite location information to terminal devices, which can include visualization screens in the area or management room, mobile terminals used by users, and the like.

[0100] In actual applications, the images collected by multiple image collection devices can be processed once, and then the processing results can be distributed to multiple terminal devices to realize information extraction and distribution. Using artificial intelligence technology, the camping density information can be extracted without manual work, effectively improving the management efficiency of management personnel and the travel experience of consumers.

[0101] Figure 8 FIG. 8 is a schematic structural block diagram of an apparatus for estimating object distribution information according to an embodiment of the present disclosure.

[0102] As shown in Figure 8 FIG. 8, the apparatus 800 for estimating object distribution information can include an information determining module 810, a second position determining module 820, and a distribution information determining module 830.

[0103] The information determining module 810 is configured to determine, for each of a plurality of images, object information of an object in the image and first position information of the object in the image, the plurality of images being obtained by an image acquisition apparatus capturing a region.

[0104] The second position determining module 820 is configured to determine, according to the first position information, second position information of the object in map data.

[0105] The distribution information determining module 830 is configured to determine, according to the second position information, estimated object distribution information of a plurality of sub-regions in the region at a target time after a current time.

[0106] According to another embodiment of the present disclosure, the distribution information determining module includes a clustering sub-module, a node generating sub-module, an edge generating sub-module, and a distribution information determining sub-module. The clustering sub-module is configured to cluster the second position information of a plurality of objects in a plurality of images in map data to obtain a plurality of clusters, each cluster corresponding to a sub-region in the region. The node generating sub-module is configured to generate a plurality of nodes in a point-edge graph, each node corresponding to a cluster. The edge generating sub-module is configured to generate edges in the point-edge graph according to the map data and the plurality of sub-regions corresponding to the plurality of clusters. The distribution information determining sub-module is configured to determine the estimated object distribution information of the plurality of sub-regions according to the point-edge graph.

[0107] According to another embodiment of the present disclosure, the attribute information of the node includes cluster center position information, the cluster center position information representing position information of the sub-region corresponding to the node in the map data. The edge generating sub-module includes an edge generating unit configured to, for any two nodes in the plurality of nodes, generate an edge between the two nodes in response to detecting that the two cluster center position information of the two nodes are on a same path in the map data and adjacent in the same path, generate an edge between the two nodes in response to detecting that the path distance between the two cluster center position information of the two nodes in the map data is equal to a distance threshold, and generate an edge between the two nodes in response to detecting that the number of object trajectories passing through the two nodes is greater than or equal to a number threshold, wherein the object trajectory is determined according to the second position information of the object at a plurality of times.

[0108] According to another embodiment of the present disclosure, the attribute information of the node includes an initial object density and historical change information, the initial object density represents an object density of a sub-region corresponding to the node at a current time, and the historical change information represents an object density change of the sub-region corresponding to the node; and the determining the sub-module according to the distribution information includes: a distribution information determining unit configured to, for each node in the point-edge graph, determine a density change amount of the node according to the historical change information of the node, and determine an estimated object density of the sub-region corresponding to the node at a target time according to the initial object density and the density change amount of the node.

[0109] According to another embodiment of the present disclosure, the number of image acquisition devices is at least two, and the collection fields of view of the at least two image acquisition devices have an intersection; the object includes a first object; and the information determining module includes a detection sub-module, a feature determining sub-module, a similarity determining sub-module, and an identification determining sub-module. The detection sub-module is configured to perform target detection on the image to obtain a detection frame for the first object in the image; the feature determining sub-module is configured to determine the feature of the first object according to the detection frame; the similarity determining sub-module is configured to determine the similarity between the feature of the first object and the features of a plurality of candidate first objects; and the identification determining sub-module is configured to determine the identification of the first object as the object information of the first object according to the similarity.

[0110] According to another embodiment of the present disclosure, the identification determining sub-module includes a first determining unit and a second determining unit. The first determining unit is configured to, in response to detecting that the maximum similarity is greater than or equal to a first similarity threshold, determine the identification of the candidate first object corresponding to the maximum similarity as the identification of the first object; and the second determining unit is configured to, in response to detecting that each similarity is less than the first similarity threshold, generate an identification other than the identifications of the plurality of candidate objects as the identification of the first object.

[0111] According to another embodiment of the present disclosure, the number of image acquisition devices is at least two, and the collection fields of view of the at least two image acquisition devices have an intersection; the object includes a second object; and the information determining module includes a deletion sub-module and an identification generating sub-module. The deletion sub-module is configured to, in response to detecting that two second objects satisfy a predetermined condition, delete one of the two second objects; and the identification generating sub-module is configured to generate the identification of the other one of the two second objects as the object information of the second object; wherein the predetermined condition includes that the distance between the two second objects in the map data is equal to a predetermined distance, and the similarity between two features of the two second objects is greater than or equal to a second similarity threshold.

[0112] In the technical solution of the present disclosure, the collection, storage, use, processing, transmission, provision, and disclosure of user personal information involved in the technical solution comply with relevant laws and regulations and do not violate public order and good customs.

[0113] In the technical solutions of the present disclosure, the authorization or consent of the user is obtained before the personal information of the user is acquired or collected.

[0114] According to an embodiment of the present disclosure, the present disclosure further provides an electronic device, comprising at least one processor; and a memory connected with the at least one processor in communication; the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method for estimating object distribution information.

[0115] According to an embodiment of the present disclosure, the present disclosure further provides a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to enable a computer to perform the method for estimating object distribution information.

[0116] According to an embodiment of the present disclosure, the present disclosure further provides a computer program product comprising a computer program, the computer program being executed by a processor to implement the method for estimating object distribution information.

[0117] Figure 9 is a structural block diagram of an electronic device for implementing the method for estimating object distribution information according to an embodiment of the present disclosure. The electronic device is intended to represent various forms of digital computers, such as laptops, desktops, tablets, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular telephones, smart phones, wearable devices, and other similar computing devices. The components shown here, their connections and relationships, and their functions, are meant to be examples only, and are not intended to limit the implementations of the present disclosure described and / or claimed in this document.

[0118] As shown in Figure 9 , the device 900 includes a computing unit 901 that can perform various appropriate actions and processes in accordance with a computer program stored in a read-only memory (ROM) 902 or a computer program loaded from a storage unit 908 into a random access memory (RAM) 903. In the RAM 903, various programs and data required for the operation of the device 900 can also be stored. The computing unit 901, the ROM 902, and the RAM 903 are connected to each other through a bus 904. An input / output (I / O) interface 905 is also connected to the bus 904.

[0119] A number of the components in device 900 are connected to I / O interface 905, including: input unit 906, such as a keyboard, mouse, etc.; output unit 907, such as various types of displays, speakers, etc.; storage unit 908, such as a disk, optical disk, etc.; and communication unit 909, such as a network card, modem, wireless communication transceiver, etc. Communication unit 909 allows device 900 to exchange information / data with other devices over computer networks, such as the Internet, and / or various telecommunication networks.

[0120] Computing unit 901 can be various general and / or special purpose processing components with processing and computing capabilities. Some examples of computing unit 901 include, but are not limited to, central processing units (CPUs), graphics processing units (GPUs), various specialized artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. Computing unit 901 performs various methods and processes described above, such as the method of estimating object distribution information. For example, in some embodiments, the method of estimating object distribution information can be implemented as a computer software program that is tangibly embodied in a machine-readable medium, such as storage unit 908. In some embodiments, part or all of the computer program can be loaded and / or installed onto device 900 via ROM 902 and / or communication unit 909. When the computer program is loaded onto RAM 903 and executed by computing unit 901, one or more steps of the method of estimating object distribution information described above can be performed. Alternatively, in other embodiments, computing unit 901 can be configured to perform the method of estimating object distribution information by any other appropriate means, such as by means of firmware.

[0121] Various implementations of the systems and techniques described above herein can be realized in digital electronic circuitry, integrated circuitry, a field programmable gate array (FPGA), an application specific integrated circuit (ASIC), a system on a chip (SOC), a complex programmable logic device (CPLD), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include implementation in one or more computer programs that are executable and / or interpretable on a programmable system including at least one programmable processor, which can be special or general purpose, coupled to receive data and instructions from, and to transmit data and instructions to, a storage system, at least one input device, and at least one output device.

[0122] Program code for carrying out methods of the present disclosure can be written in any combination of one or more programming languages. The program code can be provided to a processor or controller of a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the program code, when executed by the processor or controller, produces the functions / operations specified in the flowcharts and / or block diagrams. The program code can be executed entirely on a machine, partially on a machine, partially on a machine as a standalone software package, or entirely on a remote machine or server.

[0123] In the context of the present disclosure, a machine-readable medium can be a tangible medium that contains or stores a program for use by or in connection with an instruction execution system, apparatus, or device. The machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can include but is not limited to an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of the machine-readable storage medium will include one or more lines of electrical connections, portable computer disks, hard disk drives, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or Flash memory), optical fibers, portable compact disc read-only memories (CD-ROMs), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0124] To provide for interaction with a user, the systems and techniques described here can be implemented on a computer having a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user and a keyboard and a pointing device (e.g., a mouse or a trackball) by which the user can provide input to the computer. Other kinds of devices can be used to provide for interaction with a user as well; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form, including acoustic, speech, or tactile input.

[0125] The systems and techniques described here can be implemented in a computing system that includes a back end component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front end component (e.g., a user computer having a graphical user interface or a Web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such back end, middleware, or front end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network (LAN), a wide area network (WAN), and the Internet.

[0126] The computer system can include clients and servers. This relationship can be

[0127] It should be understood that the procedures shown above can be re-ordered, added to, or removed from, while still being within the scope of the present disclosure. For example, the steps recited in the present disclosure can be performed in parallel, in series, or in a different order, as long as the desired results of the present disclosure are achieved, and are not limited herein.

[0128] The specific embodiments described above are not intended to be limiting, and persons skilled in the art will appreciate that various modifications, combinations, sub-combinations and alternatives can be made to the specific embodiments without departing from the spirit and scope of the disclosure. Any alternatives, modifications, equivalents, and the like, along with many apparent variations that would be or become apparent to one of ordinary skill in the art are intended to be embraced by the scope of the present disclosure.

Claims

1. A method for estimating object distribution information, comprising: For each of the multiple images, determine object information of an object in the image, and determine first position information of the object in the image; The multiple images are obtained by the image acquisition device from the region; Based on the first location information, determine the second location information of the object in the map data; as well as Based on the second location information, determining the estimated object distribution information of each of the multiple sub-regions in the region at a target time after the current time includes: clustering the second location information of multiple objects in the multiple images in the map data to obtain multiple clusters, each cluster corresponding to a sub-region in the region; generating multiple nodes in the point-edge graph corresponding to the multiple clusters respectively; generating edges in the point-edge graph based on the map data and the multiple sub-regions corresponding to the multiple clusters; and determining the estimated object distribution information of each of the multiple sub-regions based on the point-edge graph.

2. The method according to claim 1, wherein, The attribute information of the node includes cluster center location information, which represents the location information of the sub-region corresponding to the node in the map data. The step of generating the edges in the point-edge graph based on the map data and the multiple sub-regions corresponding to the multiple clusters includes: For any two nodes among the plurality of nodes, In response to detecting that the location information of two cluster centers of two nodes is on the same path in the map data, and that the two cluster center location information are adjacent in the same path, an edge is generated between the two nodes; In response to detecting that the path distance between the two cluster center locations of two nodes in the map data is less than or equal to a distance threshold, an edge is generated between the two nodes; and In response to detecting that the number of object trajectories passing through two nodes is greater than or equal to a number threshold, an edge is generated between the two nodes; wherein the object trajectory is determined based on second position information of the object at multiple times.

3. The method according to claim 1, wherein, The attribute information of the node includes initial object density and historical change information. The initial object density represents the object density of the sub-region corresponding to the node at the current time, and the historical change information represents the change in object density of the sub-region corresponding to the node. Based on the point-edge graph, the estimated object distribution information of each of the multiple sub-regions is determined as follows: For each node in the aforementioned point-edge graph, Based on the historical change information of the nodes, determine the density change of the nodes; and Based on the initial object density of the node and the density change, the estimated object density of the sub-region corresponding to the node at the target time is determined.

4. The method according to claim 1, wherein, The number of image acquisition devices is at least two, and the field of view of at least two image acquisition devices overlaps; the object includes a first object; the object information for determining the object in the image includes: Perform target detection on the image to obtain a detection box for the first object in the image; Based on the detection box, the features of the first object are determined; Determine the similarity between the features of the first object and the features of multiple candidate first objects; and Based on the similarity, the identifier of the first object is determined and used as the object information of the first object.

5. The method according to claim 4, wherein, Determining the identifier of the first object based on the similarity includes: In response to detecting a maximum similarity greater than or equal to a first similarity threshold, the identifier of the candidate first object corresponding to the maximum similarity is determined as the identifier of the first object; and In response to detecting that each similarity is less than the first similarity threshold, an additional identifier other than the identifier of the plurality of candidate objects is generated as the identifier of the first object.

6. The method according to claim 1, wherein, The number of image acquisition devices is at least two, and the field of view of at least two image acquisition devices overlaps; the object includes a second object; the object information for determining the object in the image includes: In response to detecting that two second objects meet predetermined conditions, one of the two second objects is deleted; and Generate an identifier for the other of the two second objects, as the object information of the second object; The predetermined conditions include: the distance between the two second objects in the map data is less than or equal to a predetermined distance, and the similarity between two features of the two second objects is greater than or equal to a second similarity threshold.

7. An apparatus for estimating object distribution information, comprising: An information determination module is used to determine, for each of a plurality of images, object information of an object in the image, and to determine first position information of the object in the image; The multiple images are obtained by the image acquisition device from the region; The second location determination module is used to determine the second location information of the object in the map data based on the first location information; as well as The distribution information determination module is used to determine, based on the second location information, the estimated object distribution information of each of the multiple sub-regions in the region at a target time after the current time; The distribution information determination module includes: The clustering submodule is used to cluster the second location information of multiple objects in the multiple images in the map data to obtain multiple clusters, each cluster corresponding to a sub-region of the region; The node generation submodule is used to generate multiple nodes in the vertex-edge graph that correspond to the multiple clusters respectively; An edge generation submodule is used to generate edges in the point-edge graph based on the map data and multiple sub-regions corresponding to the multiple clusters; and The distribution information determination submodule is used to determine the estimated object distribution information of each of the multiple sub-regions based on the point-edge graph.

8. The apparatus according to claim 7, wherein, The attribute information of the node includes cluster center location information, which represents the location information of the sub-region corresponding to the node in the map data. The edge generation submodule includes: An edge generation unit is used to generate any two nodes from the plurality of nodes. In response to detecting that the location information of two cluster centers of two nodes is on the same path in the map data, and that the two cluster center location information are adjacent in the same path, an edge is generated between the two nodes; In response to detecting that the path distance between the two cluster center locations of two nodes in the map data is less than or equal to a distance threshold, an edge is generated between the two nodes; and In response to detecting that the number of object trajectories passing through two nodes is greater than or equal to a number threshold, an edge is generated between the two nodes; wherein the object trajectory is determined based on second position information of the object at multiple times.

9. The apparatus according to claim 7, wherein, The node's attribute information includes initial object density and historical change information. The initial object density represents the object density of the sub-region corresponding to the node at the current moment, and the historical change information represents the change in object density of the sub-region corresponding to the node. The sub-module determined based on distribution information includes: The distribution information determination unit is used for each node in the point-edge graph. Based on the historical change information of the nodes, determine the density change of the nodes; and Based on the initial object density of the node and the density change, the estimated object density of the sub-region corresponding to the node at the target time is determined.

10. The apparatus according to claim 7, wherein, The number of image acquisition devices is at least two, and the field of view of at least two image acquisition devices overlaps; the object includes a first object; the information determination module includes: The detection submodule is used to perform target detection on the image and obtain a detection box for the first object in the image; The feature determination submodule is used to determine the features of the first object based on the detection box; A similarity determination submodule is used to determine the similarity between the features of the first object and the features of multiple candidate first objects; and The identifier determination submodule is used to determine the identifier of the first object based on the similarity, and use it as the object information of the first object.

11. The apparatus according to claim 10, wherein, The sub-modules are identified based on the identifier: The first determining unit is configured to, in response to detecting that the maximum similarity is greater than or equal to a first similarity threshold, determine the identifier of the candidate first object corresponding to the maximum similarity as the identifier of the first object; and The second determining unit is configured to, in response to detecting that each similarity is less than the first similarity threshold, generate an identifier other than the identifier of the plurality of candidate objects as the identifier of the first object.

12. The apparatus according to claim 7, wherein, The number of image acquisition devices is at least two, and the acquisition fields of at least two image acquisition devices overlap; the object includes a second object; the information determination module includes: The deletion submodule is configured to delete one of the two second objects in response to detecting that two second objects meet predetermined conditions; and The identifier generation submodule is used to generate an identifier for the other of the two second objects, which serves as the object information of the second object. The predetermined conditions include: the distance between the two second objects in the map data is less than or equal to a predetermined distance, and the similarity between two features of the two second objects is greater than or equal to a second similarity threshold.

13. An electronic device, comprising: At least one processor; as well as A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1 to 6.

14. A non-transitory computer-readable storage medium storing computer instructions, wherein, The computer instructions are used to cause the computer to perform the method according to any one of claims 1 to 6.

15. A computer program product comprising a computer program that, when executed by a processor, implements the method according to any one of claims 1 to 6.

Images