CN111652160A

CN111652160A - Data processing method, device, equipment and storage medium

Info

Publication number: CN111652160A
Application number: CN202010504591.1A
Authority: CN
Inventors: 丁凯; 严石伟; 蒋楠
Original assignee: Tencent Technology Shenzhen Co Ltd
Current assignee: Tencent Technology Shenzhen Co Ltd
Priority date: 2020-06-05
Filing date: 2020-06-05
Publication date: 2020-09-11

Abstract

The application discloses a data processing method, a device, equipment and a storage medium, wherein the method comprises the following steps: acquiring first human body track data in a first type area, second human body track data in a second type area and corresponding track association information of the target area; clustering the first human body track data based on pedestrian re-identification to obtain a target clustering track; setting human identity marks for the target clustering tracks; matching the second human body trajectory data with the first human body trajectory data to determine human body trajectory matching data; determining the human identity of each human track data in the target area by using the human track matching data and the human identity of the target clustering track; and constructing the strolling track data of the person in the target area based on the human body identity of each piece of human body track data and the corresponding track correlation information. By utilizing the technical scheme provided by the embodiment of the application, the accuracy of target area management can be improved, the cost is low, and the universality is wide.

Description

Data processing method, device, equipment and storage medium

Technical Field

The present application relates to the field of artificial intelligence technologies, and in particular, to a data processing method, apparatus, device, and storage medium.

Background

With the research and progress of Artificial Intelligence (AI) technology, the Artificial Intelligence technology is being developed and applied in many fields, such as common smart retail, smart security, smart community, smart education, smart home, smart wearable devices, virtual assistants, smart speakers, smart marketing, unmanned driving, automatic driving, unmanned aerial vehicles, robots, smart medical care, smart customer service, etc. it is believed that with the development of technology, the Artificial Intelligence technology will be applied in more fields and exert more and more important value.

The application of the AI technology in the fields of intelligent retail, intelligent security, intelligent community, intelligent education and the like at present often relates to statistics of personnel filing or people flow strolling conditions of corresponding areas so as to realize the management of the areas. In the prior art, face capturing in different areas is often completed through a face camera, and techniques such as face detection and tracking are called to realize face recognition, identity profiling, people strolling condition statistics and the like based on faceID. However, in the above application, areas such as a mall, a community, and a school often include different types of sub-areas, for example, the mall includes a field area, a shop area, and the community includes a public area and a personal area, and a scheme based on a face in the prior art only achieves the situation of filing or people stream strolling for people in a certain sub-area, and cannot meet the requirements of the whole mall and the whole community for data, and the limitation is large. Therefore, there is a need to provide a more comprehensive or efficient solution.

Disclosure of Invention

The application provides a data processing method, a data processing device, data processing equipment and a storage medium, which can provide more refined and digitized people strolling information for the whole target area, improve the accuracy of target area management, and have low cost and wide universality.

In one aspect, the present application provides a data processing method, including:

acquiring a plurality of human body track data in a target area and track association information corresponding to each human body track data, wherein the human body track data comprise a plurality of first human body track data in a first type area of the target area and a plurality of second human body track data in a second type area of the target area;

clustering first human body track data generated in a preset unit time period based on pedestrian re-identification to obtain a target clustering track in the first type area;

setting a human body identity for each target clustering track in the first type area;

according to a preset frequency, matching each second human body track data with first human body track data generated in a first preset time period, and determining human body track matching data between the first type area and the second type area;

determining the human identity of each first human trajectory data and each second human trajectory data in the target area by using the human trajectory matching data and the human identity of the target clustering trajectory;

and constructing strolling track data of the person in the target area based on the human body identification of each first human body track data and each second human body track data in the target area and the corresponding track correlation information.

Another aspect provides a data processing apparatus, the apparatus comprising:

the data acquisition module is used for acquiring a plurality of human body track data in a target area and track association information corresponding to each human body track data, wherein the human body track data comprise a plurality of first human body track data in a first type area of the target area and a plurality of second human body track data in a second type area of the target area;

the clustering module is used for clustering first human body track data generated in a preset unit time period based on pedestrian re-identification to obtain a target clustering track in the first type area;

the human body identity mark setting module is used for setting a human body identity mark for each target clustering track in the first type area;

the first track matching module is used for matching each second human body track data with first human body track data generated in a first preset time period according to preset frequency, and determining human body track matching data between the first type area and the second type area;

the first human body identification determining module is used for determining the human body identification of each first human body track data and each second human body track data in the target area by using the human body track matching data and the human body identification of the target clustering track;

and the strolling track data construction module is used for constructing strolling track data of the person in the target area based on the human body identification of each first human body track data and each second human body track data in the target area and the corresponding track correlation information.

Another aspect provides a data processing apparatus comprising a processor and a memory, wherein at least one instruction or at least one program is stored in the memory, and the at least one instruction or the at least one program is loaded by the processor and executed to implement the data processing method as described above.

Another aspect provides a computer readable storage medium, in which at least one instruction or at least one program is stored, the at least one instruction or the at least one program is loaded by a processor and executed to implement the data processing method as described above.

The data processing method, the data processing device, the data processing equipment and the storage medium have the following technical effects:

the method and the device have the advantages that the human body track data in the first type area of the target area are clustered based on pedestrian re-identification to obtain the target clustering track corresponding to the human body identity, so that the building of people strolling in the first type area is realized, the building of people is performed through the human body track data, the shooting cost is low, and the universality is wide; the human body track data in the second type area of the target area is matched with the human body track data in the first type area, so that the association of the human body track data of each person in the target area among different types of areas is realized, and the communication between the human body track data in the second type area and the human body identity is further realized; and then, constructing the strolling track data of the personnel in the whole target area by combining the corresponding track correlation information. The method and the device realize that more refined and digitized people strolling information is provided for the whole target area, and the accuracy of target area management is improved.

Drawings

In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative efforts.

FIG. 1 is a schematic diagram of a data processing system provided by an embodiment of the present application;

fig. 2 is a schematic flowchart of a data processing method according to an embodiment of the present application;

fig. 3 is a schematic flowchart of a process for acquiring a plurality of human body trajectory data in a target area and trajectory related information corresponding to each human body trajectory data according to an embodiment of the present application;

FIG. 4 is a schematic flowchart illustrating a process of constructing strolling track data of a person in the target area according to an embodiment of the present application;

fig. 5 is a schematic flow chart of counting popularity in an embodiment of the present application;

FIG. 6 is a schematic flow chart of people counting according to an embodiment of the present disclosure;

fig. 7 is a schematic structural diagram of a data processing apparatus according to an embodiment of the present application;

fig. 8 is a hardware structure block diagram of an apparatus for implementing a data processing method according to an embodiment of the present application.

Detailed Description

The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only a part of the embodiments of the present application, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.

It should be noted that the terms "first," "second," and the like in the description and claims of this application and in the drawings described above are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used is interchangeable under appropriate circumstances such that the embodiments of the application described herein are capable of operation in sequences other than those illustrated or described herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or server that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.

The artificial intelligence technology is a comprehensive subject and relates to the field of extensive technology, namely the technology of a hardware level and the technology of a software level. The artificial intelligence software technology mainly comprises a computer vision technology, a voice processing technology, a natural language processing technology, machine learning/deep learning and the like. The embodiment of the specification mainly relates to a computer vision technology and deep learning in an artificial intelligence software technology.

Computer Vision technology (CV) Computer Vision is a science for researching how to make a machine "see", and further refers to that a camera and a Computer are used to replace human eyes to perform machine Vision such as identification, tracking and measurement on a target, and further image processing is performed, so that the Computer processing becomes an image more suitable for human eyes to observe or transmitted to an instrument to detect. As a scientific discipline, computer vision research-related theories and techniques attempt to build artificial intelligence systems that can capture information from images or multidimensional data. Computer vision technologies generally include image processing, image recognition, image semantic understanding, image retrieval, OCR, video processing, video semantic understanding, video content/behavior recognition, three-dimensional object reconstruction, 3D technologies, virtual reality, augmented reality, synchronous positioning, map construction, and other technologies, and also include common biometric technologies such as face recognition and fingerprint recognition.

Referring to fig. 1, fig. 1 is a schematic diagram of a data processing system according to an embodiment of the present disclosure, and as shown in fig. 1, the system may include a stream fetching module 01, a human body algorithm module 02, a message queue 03, a data distribution module 04, a background track data processing module 05, and a big data calculation module 06;

in the embodiment of the present specification, the stream fetching module 01 may be used to collect video data of people streams in areas such as a mall, a community, a school, and the like. Specifically, in the embodiment of the present specification, different image capturing devices may be selected as the stream taking module 01 in combination with an actual application scene, and specifically, the image capturing devices may include, but are not limited to, a gun camera, a dome camera, an all-in-one machine, a fisheye/panoramic camera, a pan-tilt camera, a mobile phone, a tablet, and the like.

In this embodiment of the present specification, the human body algorithm module 02 may be configured to perform corresponding calculation processing on the video data acquired by the stream taking module 01 in combination with deep learning to obtain human body trajectory data and corresponding trajectory related data, specifically, the processing may include, but is not limited to, human body detection, human body tracking, entry and exit judgment, and the like, so as to obtain multiple human body trajectory data and corresponding trajectory related data of different types of regions of the target region.

In this embodiment, the message queue 03 may be used to decouple the human body algorithm module 02 and the background trajectory data processing module 05, and may also achieve the effect of peak clipping and valley filling, so as to ensure the reliability of the system.

In the embodiment of the present specification, the data distribution module 04 is mainly configured to distribute human body trajectory data of different types of areas to trajectory data processing sub-modules of different types of areas in the background trajectory data processing module 05 according to the area type of the area where the image pickup device that collects video data is located. Specifically, for example, the area types in a shopping mall are divided into a venue area and a store area.

In this embodiment of the present specification, the background trajectory data processing module 05 may include a first trajectory data processing sub-module and a second trajectory data processing sub-module; the first trajectory data processing submodule can be used for counting people flow of the first type region, storing and registering human body trajectory data of the first type region and clustering the human body trajectory data of the first type region; the second trajectory data processing sub-module may be configured to perform retrieval matching on the acquired human body trajectory data of the second type region in the registered human body trajectory data of the first type region. In addition, the background track data processing module 05 can introduce a current limiting and automatic degradation strategy to ensure the stability and high availability of the system during the peak period.

In this embodiment, the big data calculation module 06 may be configured to perform trajectory communication between human body trajectory data in different types of areas, perform communication between human body trajectory data and human body identities, and implement determination of travel track data of a user between different types of areas in the whole area, and implement statistics of people flow and people times in different types of areas.

In some embodiments, the stream taking module 01, the human body algorithm module 02, the message queue 03, the data distribution module 04, the background track data processing module 05, and the big data calculation module 06 may be integrated in different devices, for example, the stream taking module 01 may be an independent image capturing device. The human body algorithm module 02, the message queue 03, the data distribution module 04, the background track data processing module 05, and the big data computing module 06 may be integrated in the same server, or may be integrated in different servers, and specifically, the server may include an independent physical server, a server cluster or a distributed system formed by a plurality of physical servers, or may include a cloud server providing basic cloud computing services such as cloud service, a cloud database, cloud computing, a cloud function, cloud storage, network service, cloud communication, middleware service, domain name service, security service, CDN (Content delivery network), and a big data and artificial intelligence platform.

In other embodiments, the streaming module 01 and the human body algorithm module 02 may also be integrated into the same device, and specifically, the device may include, but is not limited to, a terminal device with a camera function, such as a smartphone, a desktop computer, a tablet computer, a laptop computer, a smart speaker, a digital assistant, an Augmented Reality (AR)/Virtual Reality (VR) device, and a smart wearable device.

A data processing method according to the present application is described below, and fig. 2 is a schematic flow chart of a data processing method according to an embodiment of the present application, and the present specification provides the method operation steps as described in the embodiment or the flow chart, but more or less operation steps may be included based on conventional or non-inventive labor. The order of steps recited in the embodiments is merely one manner of performing the steps in a multitude of orders and does not represent the only order of execution. In practice, the system or server product may be implemented in a sequential or parallel manner (e.g., parallel processor or multi-threaded environment) according to the embodiments or methods shown in the figures. Specifically, as shown in fig. 2, the method may include:

s201: and acquiring a plurality of human body track data in the target area and track associated information corresponding to each human body track data.

In the embodiments of the present specification, the plurality may be at least two. In particular, the target area may include, but is not limited to, a location having a different type of area than a mall, community, school, etc. Generally, the target area may include at least two types of areas, for example, the area type of a mall may include a venue area and a store area; the area types of the community may include a public area and a personal area; schools may include field areas (playground, hallway, etc.) and indoor areas (classroom, office, etc. areas).

In an embodiment of the present specification, the body trajectory data within the target area includes a plurality of first body trajectory data within a first type area of the target area and a plurality of second body trajectory data within a second type area of the target area; specifically, the second type area may include independent areas such as a store area in a mall, a personal area in a community, and an indoor area in a school, and the first type area may include areas connecting the independent areas such as a field area in the mall, a public area in the community, and a field area in the school.

In a specific embodiment, as shown in fig. 3, the acquiring the plurality of human body trajectory data in the target area and the trajectory related information corresponding to each human body trajectory data may include:

s2011: acquiring a plurality of video data acquired in a target area and video acquisition information of each video data;

s2013: respectively extracting a first initial human body track of the first type area and a second initial human body track of the second type area based on the video data acquired in the first type area and the video data acquired in the second type area;

s2015: performing in-out behavior recognition according to each first initial human body track and the in-out shooting frame in the corresponding video acquisition information to obtain in-out behavior identification corresponding to each first initial human body track;

s2017: performing in-out behavior recognition according to each second initial human body track and the in-out shooting frame in the corresponding video acquisition information to obtain in-out behavior identification corresponding to each second initial human body track;

s2019: respectively performing characteristic purification treatment on the first initial human body track and the second initial human body track to obtain a first purified human body track and a second purified human body track;

s20111: comparing the access behavior identification corresponding to the first purified human body track of each first type area with the corresponding area type, and when the comparison result is consistent, taking the first purified human body track with the consistent comparison result as first human body track data of the first type area;

s20113: taking the second purified human body track of each second type area as second human body track data of the second type area;

s20115: taking an in-and-out behavior identifier corresponding to each first human body track data, an area identifier corresponding to video acquisition information, a shooting area type and acquisition time as track associated information of the first human body track data;

s20117: and taking the in-and-out behavior identifier corresponding to each second human body track data, the area identifier corresponding to the video acquisition information, the shooting area type and the acquisition time as track associated information of the second human body track data.

In this embodiment of the present specification, the video data may include video data collected in a first type area of the target area and video data collected in a second type area of the target area, and the video collection information includes an area identifier of a collection area of the image pickup device, a type of the image pickup area, an in-out shooting frame corresponding to the image pickup device, and a collection time (the collection time here may be a collection time of each frame of video data).

Specifically, each image capturing device acquisition region may be a sub-region in the target region, and correspondingly, the region identifier may be used to distinguish different sub-regions of the target region; specifically, the type of the camera area may represent that a human body enters or exits different camera areas, specifically, taking the target area as an area where a shopping mall is located, the type of the camera area may include, but is not limited to, an entrance, an exit, and the like. The in-out shooting frame may be used to represent that a shooting range of the image capturing apparatus belongs to a range of entering a certain area or a range of exiting the certain area, for example, the image capturing apparatus shoots facing an entrance of the certain area, and the corresponding in-out shooting frame represents that the shooting range of the image capturing apparatus belongs to the range of entering the certain area; on the contrary, the camera shooting device shoots an exit of an area, and the corresponding in-out shooting frame represents that the shooting range of the camera shooting device belongs to the range out of the area.

In this embodiment of the present description, one piece of human body trajectory data may be a trajectory of the same person acquired in an acquisition area of one piece of imaging equipment. Specifically, extracting a first initial human body trajectory of the first type region and a second initial human body trajectory of the second type region respectively based on the video data acquired in the first type region and the video data acquired in the second type region may include:

1) respectively carrying out human body detection on the video data of each subregion to obtain a plurality of human body frame snapshot data;

2) carrying out human body tracking based on the snapshot data of each human body frame to obtain original human body track data;

3) and taking the original human body track data of the first type area as a first initial human body track of the first type area, and taking the original human body track data of the second type area as a second initial human body track of the second type area.

In other embodiments, in order to improve the quality of the first initial human body trajectory and the second initial human body trajectory, after obtaining the original human body trajectory data, the method may further include:

1) acquiring the mass fraction of the snapshot data of each human body frame in each initial human body track data (a first initial human body track and a second initial human body track);

2) optimizing the initial human body trajectory data based on the mass score corresponding to each initial human body trajectory data to obtain optimized human body trajectory data;

3) and taking the optimized human body track data of the first type area as a first initial human body track of the first type area, and taking the optimized human body track data of the second type area as a second initial human body track of the second type area.

Specifically, the quality score of the human body frame snapshot data can represent the quality of the human body frame snapshot data. In the embodiment of the present specification, the quality score of the snapshot data of each human body frame can be obtained by quantifying according to the actual application requirements and the orientation, light intensity, the number of people in the same row, the degree of blur, and the like of the snapshot data of each human body frame. The preset learning model is trained by taking a large amount of human body frame snapshot data with quality score identification as training data in combination with a deep learning algorithm to obtain a quality score recognition model capable of carrying out quality score recognition on the human body frame snapshot data, and then the quality score of each human body frame snapshot data in the initial human body trajectory data can be recognized based on the quality score recognition model.

Further, the data of the human body frame snap shots with the mass score smaller than or equal to a certain threshold value or the preset number of human body frame snap shots with the smaller mass score can be removed from the initial human body track data, and optimized human body track data can be obtained.

Furthermore, for each first initial human body trajectory and each second initial human body trajectory, the corresponding in-out shooting frames can be respectively combined to perform in-out behavior recognition, and an in-out behavior identifier is obtained. Specifically, three in-and-out behavior marks, namely in-and-out behavior marks and passing in and out (passing) marks, can be determined by combining the specific trend of the initial human body trajectory (the first initial human body trajectory or the second initial human body trajectory) and the corresponding in-and-out shooting frame.

In some embodiments, the first initial human body trajectory may be taken as human body trajectory data for a first type of region (i.e., first human body trajectory data), and the second initial human body trajectory may be taken as human body trajectory data for a second type of region (second human body trajectory data).

Furthermore, in order to better ensure that each human body frame snapshot data in each human body trajectory data (first human body trajectory data or second human body trajectory data) corresponds to the same person, feature purification processing can be performed on each initial human body trajectory by combining feature data of the human body frame snapshot data. Specifically, the performing feature purification processing on the first initial human body trajectory and the second initial human body trajectory respectively to obtain a first purified human body trajectory and a second purified human body trajectory may include:

1) carrying out feature extraction on each human body snapshot data in each first initial human body track and each second initial human body track to obtain feature data of each human body snapshot data;

2) calculating the similarity between the characteristic data of each human body snapshot data;

3) and performing feature purification treatment on the corresponding first initial human body track or the second initial human body track based on the similarity to obtain a first purified human body track and a second purified human body track.

In this embodiment of the present specification, feature data of the human body frame snapshot data in the first initial human body trajectory and the second initial human body trajectory may be extracted based on a deep learning algorithm, for example, a convolutional neural network. Further, the similarity between feature data may include, but is not limited to, a euclidean distance between feature data, a cosine value, and the like.

Further, the mass score average value of the snapshot data of each human body frame in the first purified human body track after the first initial human body track purification treatment can be used as the integral mass score of the first purified human body track; and taking the mass average value of the snapshot data of each human body frame in the second purified human body track after the second initial human body track is purified as the integral mass score of the second purified human body track, so as to realize the averaging of the mass scores of the human body track data.

In this embodiment of the specification, in order to further ensure the quality of the human body trajectory data, the in-out behavior identifier corresponding to the first purified human body trajectory of the first type area may be compared with the corresponding type of the image pickup area, and when the comparison result is consistent, the first purified human body trajectory whose comparison result is consistent is used as the first human body trajectory data of the first type area. Specifically, the consistency between the access behavior identifier and the camera area type may be the consistency between the access behavior identifier and the camera area type. Specifically, taking a market scene as an example, when the entry and exit behavior identifier is entry, the corresponding shooting area type is entry, and correspondingly, the comparison result between the entry and exit behavior identifier and the shooting area type is consistent.

In practical application, the track related information is not limited to the in-out behavior identifier, the area identifier, the type of the camera shooting area and the acquisition time, but may also include track identifier information, quality score, camera shooting device identifier and line collision information (specifically, the line collision information may include a line collision frame, line collision time and a line collision sign, and the line collision frame may be an identifier frame of an entrance entering a certain area, such as an identifier frame of an entrance of a shop, a mall, and the like).

S203: and clustering first human body track data generated in a preset unit time period based on pedestrian re-identification to obtain a target clustering track in the first type area.

In this embodiment of the specification, the clustering first human body trajectory data generated in a preset unit time period based on pedestrian re-identification to obtain a target clustering trajectory in the first type region may include:

1) clustering first human body track data generated in the first preset time every second preset time in the preset unit time period based on pedestrian re-identification to obtain a temporary clustering track;

2) and clustering the temporary clustering tracks based on pedestrian re-identification every third preset time within the preset unit time period to obtain the target clustering tracks.

In the embodiment of the present specification, the preset unit time period may be preset according to practical application requirements, generally, the human body trajectory data corresponding to different clothes worn by the same person may be different, but the clothes of one person are kept unchanged within one day, and correspondingly, in order to better ensure the accuracy of the extracted human body trajectory data (corresponding to the same person), the preset unit time period may be less than or equal to 24 hours, and preferably 24 hours.

In practical application, the first human body trajectory data and the second human body trajectory data may be generated in real time, correspondingly, after the first human body trajectory data of each first type region is generated, the first human body trajectory data may be registered in a temporary clustering library, correspondingly, the newly registered first human body trajectory data in the temporary clustering library is clustered based on pedestrian re-identification every second preset time (for example, every 15 minutes), the first human body trajectory data of the same person in different first type regions are clustered together to obtain temporary clustering trajectories, correspondingly, each temporary clustering trajectory may include at least one piece of first human body trajectory data, and each temporary clustering trajectory corresponds to one temporary human body identifier.

Further, after the temporary clustering track is generated, the temporary clustering track can be registered in an accumulative clustering library; correspondingly, the new registered temporary clustering track in the cumulative clustering library can be clustered at intervals of a third preset time (for example, at every hour and every whole point) in the preset unit time period based on pedestrian re-identification, and the clustered temporary clustering track is clustered with the previously clustered temporary clustering track to obtain the target clustering track of the current first preset time point, so that the clustering of the first human body track data of the same person in different first type areas in the preset unit time period is realized.

Specifically, pedestrian Re-identification (reid) is a technique for determining whether a specific pedestrian exists in an image or a video sequence by using a computer vision technique. Correspondingly, when clustering is carried out on the human body trajectory data, whether the human body frame snapshot image (human body frame snapshot data) in the human body trajectory data corresponds to the same specific pedestrian or not can be judged through a computer vision technology.

In addition, it should be noted that the second preset time is less than the third preset time.

S205: and setting a human body identity for each target clustering track in the first type area.

In this embodiment of the present description, after obtaining target clustering tracks of the same person in different first-type regions within a preset unit time period (for example, 24 hours), a human body identity may be set for each target clustering track in the first-type region, and human body track data of different persons in the first-type region is identified, so as to implement person profiling in the first-type region.

S207: and matching each second human body track data with first human body track data generated in a first preset time period according to a preset frequency, and determining human body track matching data between the first type area and the second type area.

In the embodiment of the present specification, the track matching may be performed at a preset frequency (for example, every one hour), and the first preset time period may be a period from before a trigger time point to after the trigger time point at which the matching is performed, in consideration of the persistence of the strolling target area and the calculation performance. That is, each newly generated second human body trajectory data (the second human body trajectory data which is not retrieved and matched) is matched with all the first human body trajectory data in a period of time before and after the trigger time point, the first human body trajectory data matched with each second human body trajectory data is determined, and the human body trajectory matching data is generated. Specifically, the human body matching trajectory data may include a mapping relationship between the trajectory identification information of each second human body trajectory data and the trajectory identification information of the matched first human body trajectory data.

Specifically, the matching between the first human body trajectory data and the second human body trajectory data may be based on a deep learning algorithm, such as a convolutional neural network, to extract feature data of the first human body trajectory data and the second human body trajectory data, and calculate similarity between the feature data (the similarity between the feature data may include, but is not limited to, an euclidean distance, a cosine value, and the like between the feature data), and use the second human body trajectory data with the highest similarity as the second human body trajectory data matched with the first human body trajectory data.

In practical applications, the first human body trajectory data in the first type area may be registered to the human body search service library to determine human body trajectory data (first human body trajectory data) of the persons in a certain second type area in all the first type areas of the whole target area. Correspondingly, at the triggering time point, the newly generated second human body trajectory data can be retrieved in the human body retrieval service library, the second human body trajectory data is matched with all the first human body trajectory data in the human body retrieval service library in a period of time before and after the triggering time point, the first human body trajectory data matched with the second human body trajectory data is retrieved, and the human body trajectory matching data is generated.

In this embodiment of the present specification, through matching between the first human body trajectory data and the second human body trajectory data, association between the human body trajectory data in the first type region and the human body trajectory data in the second type region may be implemented, so as to facilitate subsequent backtracking and calculation of strolling trajectories in different types of regions of the target region.

S209: and determining the human identity of each first human body track data and each second human body track data in the target area by using the human body track matching data and the human identity of the target clustering track.

In this embodiment of the present specification, the human body matching trajectory data includes a mapping relationship between the trajectory identification information of each second human body trajectory data and the trajectory identification information of the matched first human body trajectory data. Correspondingly, the determining the human identity of each first human trajectory data and each second human trajectory data in the target area by using the human trajectory matching data and the human identity of the target clustering trajectory may include:

1) determining the human identity of each first human body track data in the target area according to the human identity of the target clustering track;

2) and determining the human identity of each second human body track data in the target area according to the human body track matching data and the human identity of each first human body track data in the target area.

In the embodiment of the present specification, a human body identification of the matched first human body trajectory data is given to each second human body trajectory data through a mapping relationship between the second human body trajectory data and the matched first human body trajectory data, so as to implement association of human body trajectory data of each person in different types of areas in a target area.

S211: and constructing strolling track data of the person in the target area based on the human body identification of each first human body track data and each second human body track data in the target area and the corresponding track correlation information.

In the embodiment of the present specification, the strolling track data may include human body track data in different types of areas corresponding to each human body identification. Specifically, a mapping relationship between each first human body trajectory data and each second human body trajectory data may be established based on the human body identity of each first human body trajectory data and each second human body trajectory data in the target area, so as to generate strolling trajectory data of the person in the target area.

In practical applications, there may be some situations where some persons make multiple round trips in the same sub-region due to some events, and accordingly, in other embodiments, as shown in fig. 4, the constructing of the strolling track data of the person in the target region based on the person identification of each first person track data and each second person track data in the target region and the corresponding track association information may include:

s2111: when the same in-out behavior identification exists in the in-out behavior identification corresponding to the first human body trajectory data of each first type region, calculating a first time interval between acquisition times corresponding to the same in-out behavior identification;

s2113: when the first time interval is smaller than or equal to a first preset threshold value, filtering non-earliest track data corresponding to the same in-out behavior identifier in the first human body track data to obtain a first filtering track;

s2115: determining the human identity identifier of the first filtering track according to the human identity identifier of each first human track data in the target area;

s2117: when the same in-out behavior identifier exists in the in-out behavior identifier corresponding to each second human body trajectory data, calculating a second time interval between acquisition times corresponding to the same in-out behavior identifier;

s2119: when the second time interval is smaller than or equal to a second preset threshold, filtering non-earliest track data corresponding to the same in-out behavior identifier in the second human body track data to obtain a second filtering track;

s21111: when the in-out behavior marks corresponding to the second filtering track in the same area comprise in-out, out-out and outdated, filtering track data with the in-out behavior marks in the second filtering track to obtain a third filtering track;

s21113: determining the human identity identifier of the third filtering track according to the human identity identifier of each second human track data in the target area;

s21115: and establishing a mapping relation between the third filtering track and the first filtering track based on the human body identification of the third filtering track and the first filtering track, and generating strolling track data of the person in the target area.

In this embodiment of the specification, the first preset threshold and the second preset threshold may be set in advance in combination with an actual application situation, taking a mall of the target area as an example, and when the type of the image pickup area corresponding to the second human body trajectory data is an incoming store or an outgoing store, the second preset threshold may be 60 s; when the type of the image capture area corresponding to the first human body trajectory data is an entrance, an exit, an entrance, and an exit, the first preset threshold may be 30s, and when the type of the image capture area corresponding to the first human body trajectory data is a special area (such as a exhibition stand, etc.), the first preset threshold may be 60 s.

In a specific embodiment, assuming that the in-out behavior identifier corresponding to a certain body trajectory data (the first body trajectory data or the second body trajectory data) includes in-out, the corresponding non-earliest trajectory data may be trajectory data generated according to the acquisition time and later in time.

In the embodiment of the specification, the first human body trajectory data and the second human body trajectory data are filtered, so that identity-based deduplication can be realized, and the validity of the human body trajectory data between different types of areas subjected to correlation can be better ensured.

In other embodiments, in order to further grasp the strolling situation of people in the target area such as the whole shopping mall, school, community, etc., statistics may be further performed on the people who flow in different types of areas, and accordingly, as shown in fig. 5, the method further includes:

s501: when the same in-out behavior identification exists in the in-out behavior identification corresponding to the first human body trajectory data of each first type region, calculating a first time interval between acquisition times corresponding to the same in-out behavior identification;

s503: when the first time interval is smaller than or equal to a first preset threshold value, filtering non-earliest track data corresponding to the same in-out behavior identifier in the first human body track data to obtain a first filtering track;

s505: and determining the number of people in the first type area according to the type of the image pickup area corresponding to the first filtering track and the area identification.

In this embodiment of the present specification, the popularity of people in the first type area may include at least one of the popularity of people in each first type area, the popularity of people out of each first type area, the popularity of people in the first type area of the target area, and the popularity of people out of the first type area of the target area.

Specifically, the behavior identifier may be generated according to the type of the image capturing area, and different in-out or out-of events corresponding to the type of the image capturing area are corresponding to different behavior identifiers, for example, the type of the image capturing area corresponding to a certain first filtering track is in-out, and correspondingly, in-out and out-of may correspond to two different event identifiers.

Furthermore, the number of the first filtering tracks marked as corresponding events of different first type areas is accumulated by combining the area marks, and the accumulated number is used as the number of people entering the first type area of the target area; in addition, the number of the first filtering tracks marked by the corresponding events of different first type areas is accumulated, and the accumulated number is used as the number of people in the first type area of the target area.

Further, accumulating the number of the first filtering tracks marked as incoming corresponding to the event mark of each first type area by combining the area marks, and taking the accumulated number as the number of people flowing in the first type area; in addition, the number of the first filtering tracks marked by the event marks corresponding to the area marks of each first type area is accumulated, and the accumulated number is used as the number of people flowing in the first type area.

S507: when the same in-out behavior identifier exists in the in-out behavior identifier corresponding to each second human body trajectory data, calculating a second time interval between acquisition times corresponding to the same in-out behavior identifier;

s509: when the second time interval is smaller than or equal to a second preset threshold, filtering non-earliest track data corresponding to the same in-out behavior identifier in the second human body track data to obtain a second filtering track;

s511: when the in-out behavior marks corresponding to the second filtering track in the same area comprise in-out, out-out and outdated, filtering track data with the in-out behavior marks in the second filtering track to obtain a third filtering track;

s513: and determining the people flow and people times of the second type area according to the in-out behavior identification and the area identification corresponding to the third filtering track.

In this embodiment of the present specification, the popularity ranking of people in the second type area may include at least one of the popularity ranking of people in each second type area, the popularity ranking of people out of each second type area, the popularity ranking of people in the second type area of the target area, and the popularity ranking of people out of the second type area of the target area.

Specifically, the event identifier may be generated according to the type of the image capturing area, and different in-out or out-of events corresponding to the type of the image capturing area are corresponding to different event identifiers, for example, the type of the image capturing area corresponding to a certain third filtering track is in-out, and correspondingly, in-out and out-of may correspond to two different event identifiers.

Furthermore, the number of the third filtering tracks marked as corresponding events of different second type areas is accumulated by combining the area marks, and the accumulated number is used as the number of people entering the second type area of the target area; in addition, the number of the third filtering tracks marked by the corresponding events of the different second type areas is accumulated, and the accumulated number is used as the number of people in the second type area of the target area.

Further, accumulating the number of third filtering tracks marked as incoming by the event identifier corresponding to the area identifier of each second type area by combining the area identifiers, and taking the accumulated number as the number of incoming people in the second type area; in addition, the number of the third filtering tracks marked by the event marks corresponding to the area marks of each second type area is accumulated, and the accumulated number is used as the number of people flowing in the second type area.

In addition, the filtering processing on the trajectory data in the embodiment of performing pedestrian counting and the filtering processing on the trajectory data in the embodiment of constructing the strolling trajectory data are both used for filtering repeated trajectories, and accordingly, the obtained first filtering trajectory, the obtained second filtering trajectory and the obtained third filtering trajectory are consistent.

In other embodiments, in order to reflect the people flow more accurately, the number of people flow in the target area may be counted, and accordingly, as shown in fig. 6, the method may further include:

s601: determining the human identity of the first filtering track based on the human identity of the target clustering track;

in this embodiment of the present specification, since the target clustering trajectory includes one or more first human body trajectory data, correspondingly, different first filtering trajectories may correspond to a human body id.

S603: and determining the number of people in the first type area according to the human identity identification of the first filtering track, the corresponding shooting area type and the area identification.

In this embodiment, the number of people in the first type area may include at least one of the number of people in each first type area, the number of people out of each first type area, the number of people in the first type area of the target area, and the number of people out of the first type area of the target area.

Specifically, the event identifier may be generated according to the type of the image capturing area, and different in-out or out-of events corresponding to the type of the image capturing area are mapped to different event identifiers, for example, the type of the image capturing area corresponding to a certain first filtering track is in-out, and correspondingly, in-out and out-of may correspond to two different event identifiers.

Further, the number of the human body identification marks of the first filtering track corresponding to the event marks of different first type areas can be accumulated by combining the area marks, and the accumulated number is used as the number of people in the first type area of the target area; in addition, the number of the human body identification marks of the first filtering track marked by the corresponding events of different first type areas is accumulated, and the accumulated number is used as the number of people in the first type area of the target area.

Further, the number of the human body identification marks of the first filtering track corresponding to the event mark of each first type area can be accumulated by combining the area marks, and the accumulated number is used as the number of people in the first type area; in addition, the number of the human body identification marks of the first filtering track marked by the event mark corresponding to the area mark of each first type area is accumulated, and the accumulated number is used as the number of people in the first type area.

In addition, it should be noted that, in some scenarios, in order to ensure the real-time performance of people counting, the temporary human body identifier of the corresponding temporary clustering track may be directly used as the human body identity identifier of the first filtering track.

S605: matching each second human body trajectory data with first human body trajectory data in a third preset time period before the second human body trajectory data generation time, and determining first human body trajectory data matched with the second human body trajectory data;

s607: determining the human identity of the matched first human body track data according to the human identity of the target clustering track;

s609: determining the human identity identifier of the third filtering track according to the human identity identifier of the matched first human trajectory data;

s611: and determining the people flow number in the second type area according to the human body identity identification of the third filtering track, the corresponding in-out behavior identification and the area identification.

In this embodiment, the number of people in the second type area may include at least one of the number of people in each second type area, the number of people in the second type area of the target area, and the number of people in the second type area of the target area.

Further, the number of the human body identification marks of the third filtering track corresponding to the event marks in different second type areas can be accumulated by combining the area marks, and the accumulated number is used as the number of people in the second type area of the target area; in addition, the number of the human body identification marks of the third filtering track marked by the corresponding events of different second type areas is accumulated, and the accumulated number is used as the number of people in the second type area of the target area.

Furthermore, the number of the human body identification marks of the third filtering track marked by the event corresponding to each second type area can be accumulated by combining the area marks, and the accumulated number is used as the number of people entering the second type area; in addition, the number of the human body identification marks of the third filtering track marked by the event mark corresponding to each second type area is accumulated, and the accumulated number is used as the number of people in the second type area.

In addition, it should be noted that, in some scenarios, in order to ensure the real-time performance of people counting, the temporary human body identifier of the corresponding temporary clustering track may be directly used as the human body identity identifier of the third filtering track.

In other embodiments, when the target area is a mall, a scheme of attendance checking by store personnel can be added to better perfect management of the whole mall, and correspondingly, the face data of the store personnel can be collected; adding the face data into an employee library; correspondingly, when a clerk punches a card every time when working, the clerk can punch the card by swiping the face, and particularly, after one face data is collected, the face data of the clerk in the employee library can be called to verify, so that the card punching operation is realized.

According to the technical scheme provided by the embodiment of the specification, the human body track data in the first type area of the target area is clustered based on pedestrian re-identification in the specification to obtain the target clustering track corresponding to the human body identity, so that the filing of people strolling in the first type area is realized, the filing of people is performed through the human body track data, the shooting cost is low, and the universality is wide; the human body track data in the second type area of the target area is matched with the human body track data in the first type area, so that the association of the human body track data of each person in the target area among different types of areas is realized, and the communication between the human body track data in the second type area and the human body identity is further realized; and then, constructing the strolling track data of the personnel in the whole target area by combining the corresponding track correlation information. The method and the device realize that more refined and digitized people strolling information is provided for the whole target area, and the accuracy of target area management is improved. In addition, on the basis of constructing the strolling track data of the personnel in the whole target area, the people flow times and the people flow number in different areas are counted, so that the people flow condition of the target area is more comprehensively known, and powerful data support is provided for the follow-up management of the whole target area.

In a specific application scenario, for example, in management of a shopping mall, the technical scheme provided by the embodiment of the specification can be used for clustering the human body track data in the area based on pedestrian re-identification to obtain a target clustering track corresponding to the human body identity, so as to realize filing of users visiting in the area of the shopping mall, and realize association of the human body track data between the area of each user in the target area through matching of the human body track data in the area of the shopping mall and the human body track data in the area of the shopping mall, so as to realize communication between the human body track data in the area of the shopping mall and the human body identity; and then, constructing the strolling track data of the personnel in the whole shopping mall by combining the corresponding track correlation information. More refined and digitized passenger flow information is provided for the shopping malls and the shops, so that the shopping malls and the shops can make more reasonable operation decisions according to the shopping behaviors and the passenger flow changes of customers, and the service income is improved. The user is built through the human body track data, the camera shooting cost is low, the universality is wide, and the system can be applied to various current market scenes, particularly non-standard scenes such as difficult erection of a gun camera and the like or markets sensitive to the cost of the gun camera; in addition, on the basis of constructing the strolling track data of personnel in the whole shopping mall, statistics of the number of people in different areas and the number of people in the different areas is realized, so that people flow conditions of the shopping mall can be more comprehensively known, and powerful data support is provided for subsequent analysis of consumer consumption habits, important attention, areas of stay, shop planning, storefront renting, passenger group diversion, shop-level goods display optimization, shop address selection, purchase, sale and storage and other management.

An embodiment of the present application further provides a data processing apparatus, as shown in fig. 7, the apparatus includes:

a data obtaining module 710, configured to obtain a plurality of human body trajectory data in a target area and trajectory related information corresponding to each human body trajectory data, where the human body trajectory data includes a plurality of first human body trajectory data in a first type area of the target area and a plurality of second human body trajectory data in a second type area of the target area;

the clustering module 720 may be configured to cluster first human trajectory data generated in a preset unit time period based on pedestrian re-identification to obtain a target clustering trajectory in the first type region;

a human identity setting module 730, configured to set a human identity for each target clustering track in the first type region;

the first trajectory matching module 740 may be configured to match each second human body trajectory data with first human body trajectory data generated within a first preset time period according to a preset frequency, and determine human body trajectory matching data between the first type area and the second type area;

a first human body identification determination module 750, configured to determine a human body identification of each first human body trajectory data and each second human body trajectory data in the target area by using the human body trajectory matching data and the human body identification of the target clustering trajectory;

the strolling track data building module 760 may be configured to build strolling track data of a person in the target area based on the human identity and the corresponding track association information of each first human body track data and each second human body track data in the target area.

In some embodiments, the clustering module 720 may include:

the first clustering unit is used for clustering first human body track data generated in first preset time every second preset time in the preset unit time period based on pedestrian re-identification to obtain a temporary clustering track;

and the second clustering unit is used for clustering the temporary clustering track based on pedestrian re-identification at intervals of a third preset time within the preset unit time period to obtain the target clustering track.

In some embodiments, the data acquisition module 710 may include:

the data acquisition unit is used for acquiring a plurality of video data acquired in a target area and video acquisition information of each video data, wherein the video data comprises video data acquired in a first type area of the target area and video data acquired in a second type area of the target area, and the video acquisition information comprises an area identifier of an acquisition area of the camera equipment, a camera area type, a shooting frame for the camera equipment to go in and out and acquisition time;

the human body track extraction unit is used for respectively extracting a first initial human body track of the first type area and a second initial human body track of the second type area based on the video data acquired in the first type area and the video data acquired in the second type area;

the first in-out behavior recognition unit is used for recognizing in-out behaviors according to each first initial human body track and the in-out shooting frame in the corresponding video acquisition information to obtain an in-out behavior identifier corresponding to each first initial human body track;

the second in-out behavior recognition unit is used for recognizing in-out behaviors according to each second initial human body track and the in-out shooting frame in the corresponding video acquisition information to obtain an in-out behavior identifier corresponding to each second initial human body track;

the characteristic purification processing unit is used for respectively carrying out characteristic purification processing on the first initial human body track and the second initial human body track to obtain a first purified human body track and a second purified human body track;

the comparison unit is used for comparing the access behavior identification corresponding to the first purified human body track of each first type area with the corresponding camera area type;

a first human body trajectory data determining unit configured to, when the comparison result is identical, take a first purified human body trajectory, of which the comparison result is identical, as first human body trajectory data of the first type region;

the second human body track data determining unit is used for taking the second purified human body track of each second type area as the second human body track data of the second type area;

the first track associated information determining unit is used for taking an in-and-out behavior identifier corresponding to each first human body track data, an area identifier in corresponding video acquisition information, a shooting area type and acquisition time as track associated information of the first human body track data;

and the second track associated information determining unit is used for taking the in-and-out behavior identifier corresponding to each second human body track data, the area identifier corresponding to the video acquisition information, the shooting area type and the acquisition time as the track associated information of the second human body track data.

In some embodiments, the apparatus further comprises:

the first time interval calculation module is used for calculating a first time interval between acquisition times corresponding to the same in-out behavior identification when the same in-out behavior identification exists in the in-out behavior identification corresponding to the first human body trajectory data of each first type region;

a first filtering module, configured to, when the first time interval is less than or equal to a first preset threshold, filter non-earliest trajectory data corresponding to the same in-and-out behavior identifier in the first human trajectory data to obtain a first filtered trajectory;

the first people stream people number determining module is used for determining the people stream people number of the first type area according to the shooting area type and the area identification corresponding to the first filtering track;

the second time interval calculation module is used for calculating a second time interval between acquisition times corresponding to the same in-out behavior identification when the same in-out behavior identification exists in the in-out behavior identification corresponding to each second human body trajectory data;

the second filtering module is used for filtering non-earliest track data corresponding to the same in-out behavior identifier in the second human body track data to obtain a second filtering track when the second time interval is smaller than or equal to a second preset threshold;

the third filtering module is used for filtering the track data with the access behavior identifier in the second filtering track when the access behavior identifier corresponding to the second filtering track in the same area comprises an access behavior identifier, an exit behavior identifier and an outdated behavior identifier to obtain a third filtering track;

and the second people stream and people number determining module is used for determining the people stream and people number of the second type area according to the entrance and exit behavior identifier and the area identifier corresponding to the third filtering track.

In some embodiments, the apparatus further comprises:

the second human identity identification determining module is used for determining the human identity identification of the first filtering track based on the human identity identification of the target clustering track;

the first people stream number determining module is used for determining the people stream number in the first type area according to the human body identity identification of the first filtering track, the corresponding shooting area type and the area identification;

the second track matching module is used for matching each second human body track data with first human body track data in a third preset time period before the second human body track data generation time, and determining first human body track data matched with the second human body track data;

the third human body identity identification determining module is used for determining the human body identity identification of the matched first human body track data according to the human body identity identification of the target clustering track;

the fourth human body identification determining module is used for determining the human body identification of the third filtering track according to the human body identification of the matched first human body track data;

and the second people flow number determining module is used for determining the people flow number of the second type area according to the human body identity identification of the third filtering track, the corresponding in-out behavior identification and the area identification.

In some embodiments, the first human identity identification determination module 750 may include:

the first human body identification determining unit is used for determining the human body identification of each first human body track data in the target area according to the human body identification of the target clustering track;

and the second human body identity identification determining unit is used for determining the human body identity identification of each second human body track data in the target area according to the human body track matching data and the human body identity identification of each first human body track data in the target area.

In some embodiments, the strolling track data construction module 760 includes:

the first time interval calculation unit is used for calculating a first time interval between acquisition times corresponding to the same in-out behavior identification when the same in-out behavior identification exists in the in-out behavior identification corresponding to the first human body trajectory data of each first type region;

the first filtering unit is used for filtering non-earliest trajectory data corresponding to the same in-out behavior identifier in the first human body trajectory data to obtain a first filtering trajectory when the first time interval is smaller than or equal to a first preset threshold;

the third human body identification mark determining unit is used for determining the human body identification mark of the first filtering track according to the human body identification mark of each first human body track data in the target area;

the second time interval calculation unit is used for calculating a second time interval between acquisition times corresponding to the same in-out behavior identification when the same in-out behavior identification exists in the in-out behavior identification corresponding to each second human body trajectory data;

the second filtering unit is used for filtering non-earliest track data corresponding to the same in-out behavior identifier in the second human body track data to obtain a second filtering track when the second time interval is smaller than or equal to a second preset threshold;

the third filtering unit is used for filtering the track data with the access behavior identifier in the second filtering track when the access behavior identifier corresponding to the second filtering track in the same area comprises access, exit and outdated to obtain a third filtering track;

a fourth human body identification determining unit, configured to determine a human body identification of the third filtering track according to the human body identification of each second human body track data in the target area;

and the strolling track data generating unit is used for establishing a mapping relation between the third filtering track and the first filtering track based on the human body identification of the third filtering track and the first filtering track, and generating strolling track data of the person in the target area.

The device and method embodiments in the device embodiment are based on the same application concept.

The embodiment of the present application provides a data processing device, which includes a processor and a memory, where the memory stores at least one instruction or at least one program, and the at least one instruction or the at least one program is loaded and executed by the processor to implement the data processing method provided by the above method embodiment.

The memory may be used to store software programs and modules, and the processor may execute various functional applications and data processing by operating the software programs and modules stored in the memory. The memory can mainly comprise a program storage area and a data storage area, wherein the program storage area can store an operating system, application programs needed by functions and the like; the storage data area may store data created according to use of the apparatus, and the like. Further, the memory may include high speed random access memory, and may also include non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other volatile solid state storage device. Accordingly, the memory may also include a memory controller to provide the processor access to the memory.

The method provided by the embodiment of the application can be executed in a mobile terminal, a computer terminal, a server or similar operation equipment. Fig. 8 is a hardware structure block diagram of an apparatus for implementing a data processing method according to an embodiment of the present application. As shown in fig. 8, the apparatus 800 may have a relatively large difference due to different configurations or performances, and may include one or more Central Processing Units (CPUs) 810 (the processor 810 may include but is not limited to a Processing device such as a microprocessor MCU or a programmable logic device FPGA), a memory 830 for storing data, one or more storage media 820 (e.g., one or more mass storage devices) for storing applications 823 or data 822. Memory 830 and storage medium 820 may be, among other things, transient or persistent storage. The program stored in storage medium 820 may include one or more modules, each of which may include a series of instruction operations for a server. Still further, central processor 810 may be configured to communicate with storage medium 820 to perform a series of instruction operations in storage medium 820 on device 800. Apparatus 800 may also include one or more power supplies 860, one or more wired or wireless network interfaces 850, one or more outputsAn input/output interface 840, and/or one or more operating systems 821, such as Windows Server^TM，Mac OS X^TM，Unix^TM,Linux^TM，FreeBSD^TMAnd so on.

The input-output interface 840 may be used to receive or transmit data via a network. Specific examples of such networks may include wireless networks provided by the communication provider of the device 800. In one example, i/o Interface 840 includes a Network adapter (NIC) that may be coupled to other Network devices via a base station to communicate with the internet. In one example, the input/output interface 840 may be a Radio Frequency (RF) module, which is used for communicating with the internet in a wireless manner.

It will be understood by those skilled in the art that the structure shown in fig. 8 is only an illustration and is not intended to limit the structure of the electronic device. For example, device 800 may also include more or fewer components than shown in FIG. 8, or have a different configuration than shown in FIG. 8.

Embodiments of the present application further provide a storage medium, which may be disposed in a device to store at least one instruction related to implementing a data processing method in the method embodiments, or at least one program, where the at least one instruction or the at least one program is loaded and executed by the processor to implement the data processing method provided in the method embodiments.

Alternatively, in this embodiment, the storage medium may be located in at least one network server of a plurality of network servers of a computer network. Optionally, in this embodiment, the storage medium may include, but is not limited to: a U-disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a removable hard disk, a magnetic disk, or an optical disk, which can store program codes.

According to the data processing method, the device, the equipment or the storage medium, the human body track data in the first type area are clustered based on pedestrian re-identification to obtain the target clustering track corresponding to the human body identity, so that the profiling of people visiting in the first type area is realized, the profiling of people is performed through the human body track data, the shooting cost is low, and the universality is wide; the human body track data in the second type area is matched with the target clustering track in the first type area, so that the association of the human body track data of each person in the target area among different types of areas is realized, and the communication between the human body track data in the second type area and the human body identity is further realized; and then, constructing the strolling track data of the personnel in the whole target area by combining the corresponding track correlation information. The method and the device realize that more refined and digitized people strolling information is provided for the whole target area, and the accuracy of target area management is improved. In addition, on the basis of constructing the strolling track data of the personnel in the whole target area, the people flow times and the people flow number in different areas are counted, so that the people flow condition of the target area is more comprehensively known, and powerful data support is provided for the follow-up management of the whole target area.

It should be noted that: the sequence of the embodiments of the present application is only for description, and does not represent the advantages and disadvantages of the embodiments. And specific embodiments thereof have been described above. Other embodiments are within the scope of the following claims. In some cases, the actions or steps recited in the claims may be performed in a different order than in the embodiments and still achieve desirable results. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing may also be possible or may be advantageous.

The embodiments in the present specification are described in a progressive manner, and the same and similar parts among the embodiments are referred to each other, and each embodiment focuses on the differences from the other embodiments. In particular, for the apparatus, device and storage medium embodiments, since they are substantially similar to the method embodiments, the description is relatively simple, and reference may be made to some descriptions of the method embodiments for relevant points.

It will be understood by those skilled in the art that all or part of the steps for implementing the above embodiments may be implemented by hardware, or may be implemented by a program instructing relevant hardware to implement the above embodiments, where the program may be stored in a computer-readable storage medium, and the above-mentioned storage medium may be a read-only memory, a magnetic disk, an optical disk, or the like.

The above description is only exemplary of the present application and should not be taken as limiting the present application, as any modification, equivalent replacement, or improvement made within the spirit and principle of the present application should be included in the protection scope of the present application.

Claims

1. A method of data processing, the method comprising:

2. The method according to claim 1, wherein the clustering first human trajectory data generated in a preset unit time period based on pedestrian re-identification to obtain a target clustering trajectory in the first type region comprises:

clustering first human body track data generated in the first preset time every second preset time in the preset unit time period based on pedestrian re-identification to obtain a temporary clustering track;

and clustering the temporary clustering tracks based on pedestrian re-identification every third preset time within the preset unit time period to obtain the target clustering tracks.

3. The method according to claim 1, wherein the obtaining of the plurality of human body trajectory data in the target area and the trajectory related information corresponding to each human body trajectory data comprises:

acquiring a plurality of video data acquired in a target area and video acquisition information of each video data, wherein the video data comprises video data acquired in a first type area of the target area and video data acquired in a second type area of the target area, and the video acquisition information comprises area identification of an acquisition area of camera equipment, type of the camera area, a shooting frame for the camera equipment to go in and out and acquisition time;

respectively extracting a first initial human body track of the first type area and a second initial human body track of the second type area based on the video data acquired in the first type area and the video data acquired in the second type area;

performing in-out behavior recognition according to each first initial human body track and the in-out shooting frame in the corresponding video acquisition information to obtain in-out behavior identification corresponding to each first initial human body track;

performing in-out behavior recognition according to each second initial human body track and the in-out shooting frame in the corresponding video acquisition information to obtain in-out behavior identification corresponding to each second initial human body track;

respectively performing characteristic purification treatment on the first initial human body track and the second initial human body track to obtain a first purified human body track and a second purified human body track;

comparing the access behavior identification corresponding to the first purified human body track of each first type area with the corresponding camera shooting area type, and when the comparison result is consistent, taking the first purified human body track with the consistent comparison result as first human body track data of the first type area;

taking the second purified human body track of each second type area as second human body track data of the second type area;

taking an in-and-out behavior identifier corresponding to each first human body track data, an area identifier corresponding to video acquisition information, a shooting area type and acquisition time as track associated information of the first human body track data;

and taking the in-and-out behavior identifier corresponding to each second human body track data, the area identifier corresponding to the video acquisition information, the shooting area type and the acquisition time as track associated information of the second human body track data.

4. The method of claim 3, further comprising:

when the same in-out behavior identification exists in the in-out behavior identification corresponding to the first human body trajectory data of each first type region, calculating a first time interval between acquisition times corresponding to the same in-out behavior identification;

when the first time interval is smaller than or equal to a first preset threshold value, filtering non-earliest track data corresponding to the same in-out behavior identifier in the first human body track data to obtain a first filtering track;

determining the number of people flowing in the first type area according to the type and the area identification of the camera area corresponding to the first filtering track;

when the same in-out behavior identifier exists in the in-out behavior identifier corresponding to each second human body trajectory data, calculating a second time interval between acquisition times corresponding to the same in-out behavior identifier;

when the second time interval is smaller than or equal to a second preset threshold, filtering non-earliest track data corresponding to the same in-out behavior identifier in the second human body track data to obtain a second filtering track;

when the in-out behavior marks corresponding to the second filtering track in the same area comprise in-out, out-out and outdated, filtering track data with the in-out behavior marks in the second filtering track to obtain a third filtering track;

and determining the people flow and people times of the second type area according to the in-out behavior identification and the area identification corresponding to the third filtering track.

5. The method of claim 4, further comprising:

determining the human identity of the first filtering track based on the human identity of the target clustering track;

determining the number of people in the first type area according to the human identity identification of the first filtering track, the corresponding shooting area type and the area identification;

matching each second human body trajectory data with first human body trajectory data in a third preset time period before the second human body trajectory data generation time, and determining first human body trajectory data matched with the second human body trajectory data;

determining the human identity of the matched first human body track data according to the human identity of the target clustering track;

determining the human identity identifier of the third filtering track according to the human identity identifier of the matched first human trajectory data;

and determining the people flow number in the second type area according to the human body identity identification of the third filtering track, the corresponding in-out behavior identification and the area identification.

6. The method of claim 1, wherein the determining the human identity of each first human trajectory data and each second human trajectory data within the target area using the human trajectory matching data and the human identity of the target clustering trajectory comprises:

determining the human identity of each first human body track data in the target area according to the human identity of the target clustering track;

and determining the human identity of each second human body track data in the target area according to the human body track matching data and the human identity of each first human body track data in the target area.

7. The method of claim 3, wherein the constructing strolling track data of the person in the target area based on the human identity and the corresponding track correlation information of each first human track data and each second human track data in the target area comprises:

determining the human identity identifier of the first filtering track according to the human identity identifier of each first human track data in the target area;

determining the human identity identifier of the third filtering track according to the human identity identifier of each second human track data in the target area;

and establishing a mapping relation between the third filtering track and the first filtering track based on the human body identification of the third filtering track and the first filtering track, and generating strolling track data of the person in the target area.

8. A data processing apparatus, characterized in that the apparatus comprises:

9. A data processing apparatus, characterized in that the apparatus comprises a processor and a memory, in which at least one instruction or at least one program is stored, which is loaded and executed by the processor to implement the data processing method according to any one of claims 1 to 7.

10. A computer-readable storage medium, in which at least one instruction or at least one program is stored, which is loaded and executed by a processor to implement the data processing method according to any one of claims 1 to 7.