Supervision object positioning and tracking method, system and equipment based on video analysis and medium

Abstract

The invention provides a supervised object positioning and tracking method, system and device based on video analysis and a medium, and relates to the technical field of video analysis, positioning and tracking, and the method comprises the steps: obtaining a supervised object video through a camera; obtaining data information of a supervision place, and constructing a digital twinborn model of the supervision place; processing the supervision object mobile video to obtain first supervision object information; on the basis of a convolutional neural network architecture, identifying a supervision object in video monitoring of the supervision place to obtain second supervision object information; according to the behavior recognition model, performing real-time analysis on the behavior of the supervised object by using a deep learning method; and judging a result, if an abnormal event is detected, giving an alarm, and performing remote control on equipment in the supervision place through the system. The method has the beneficial effects that the track of the supervisor can be better tracked according to the neural network model and the behavior recognition model, so that the supervisor is helped to determine position information possibly positioned and tracked, track and position the position of the supervisor in real time and display the position information in the digital twin model, and the dynamic state of the supervisor is monitored in real time.

Description

Technical Field

[0001] The present invention relates to the technical field of video analysis positioning and tracking, and in particular to a method, system, device and medium for positioning and tracking a supervised object based on video analysis. Background Art

[0002] With the development of police informationization, many cities in China have high-precision cameras on major roads and streets to monitor traffic information. The existence of cameras has played a very important role in some criminal investigations. However, in many cases, due to the large amount of video information in video surveillance, observing the whereabouts of the supervised object with the naked eye will not only take a long time, but may also cause some details to be missed. Even if the dynamics of the supervised object are observed manually, it is time-consuming and laborious, and it is impossible to monitor the whereabouts of the supervised object in real time, analyze the abnormal behavior of the supervised object, and control the abnormal behavior of the supervised object in the bud, and thus it is impossible to take effective corresponding measures. Summary of the invention

[0003] The purpose of the present invention is to provide a method, system, device and medium for positioning and tracking a supervised object based on video analysis to improve the above problems. In order to achieve the above purpose, the technical solution adopted by the present invention is as follows:

[0004] In a first aspect, the present application provides a method for positioning and tracking a supervised object based on video analysis, comprising:

[0005] Obtaining videos of supervised objects through cameras, wherein the supervised object movement videos include supervised object movement clips, supervised object action clips, and supervised object activity distribution;

[0006] Obtaining data information of the supervision site, where the data information includes spatial information, facility information, equipment information and supervision object information of the supervision site, and building a digital twin model of the supervision site based on the data information of the supervision site;

[0007] Processing the mobile video of the supervised object to obtain processed first supervised object information; and based on the convolutional neural network architecture, inputting the first supervised object information into the digital twin model for training, identifying the supervised object in the video surveillance of the supervised place, and obtaining second supervised object information, wherein the second supervised object information includes the location information of each supervised object and the identity information corresponding to each supervised object;

[0008] Based on the behavior recognition model, the deep learning method is used to analyze the behavior of the supervised object in real time, where the behavior of the supervised object includes the detection and classification of abnormal behavior, and the analysis results are obtained;

[0009] The results of the analysis are judged. If an abnormal event is detected, an alarm is issued and the equipment in the supervision place is remotely controlled through the system. If no abnormal event is detected, the supervision object will continue to be located and tracked.

[0010] Preferably, the mobile video of the supervised object is processed to obtain the processed first supervised object information, which includes:

[0011] Perform video segmentation on the moving video of the supervised object, wherein the segmentation process is to extract a segment every five seconds, and each segment includes a photo of the supervised object's activities;

[0012] Extract voice data and time-series-based image data from the clips using editing software; segment the voice data using Bayesian theorem to obtain voice clips;

[0013] Preprocessing the image data, wherein the preprocessing process includes processing the image data using Gaussian filtering, and optimizing the image data after the noise is weakened according to the edge loss function to obtain image fragments;

[0014] The voice segment and the image segment are fused and aggregated to obtain processed first supervision object information.

[0015] Preferably, the obtaining of the second supervision object information includes:

[0016] Extracting the area range of the motion change of the supervision object target corresponding to the second supervision object information as the target template;

[0017] Superimpose the target template and the real-time image location, and calculate the corresponding value of the similarity measure between the two;

[0018] The position corresponding to the maximum similarity in the corresponding values ​​is extracted, and then the position of the target of the supervision object is obtained;

[0019] The location of the supervised object target is locked, and the supervised object information identity is identified based on the preset information database.

[0020] Preferably, the behavior recognition model includes a classifier, a feature extractor and a decoder, and the behavior recognition model is optimized using a back propagation algorithm and an optimizer, a cross entropy loss function is selected, the parameters of the behavior recognition model are updated, and multiple iterations of training and model tuning are performed until the behavior recognition model converges and achieves the matching performance.

[0021] In a second aspect, the present application also provides a supervision object positioning and tracking system based on video analysis, including a first acquisition module, a second acquisition module, a processing module, an analysis module and a judgment module, wherein:

[0022] The first acquisition module is used to acquire the video of the supervised object through the camera, wherein the supervised object movement video includes the supervised object movement segment, the supervised object action segment and the supervised object activity distribution;

[0023] The second acquisition module is used to acquire data information of the supervision site, where the data information includes spatial information, facility information, equipment information and supervision object information of the supervision site, and to build a digital twin model of the supervision site based on the data information of the supervision site;

[0024] Processing module: used to process the mobile video of the supervised object to obtain the processed first supervised object information; and based on the convolutional neural network architecture, input the first supervised object information into the digital twin model for training, identify the supervised object in the video surveillance of the supervised place, and obtain the second supervised object information, wherein the second supervised object information includes the location information of each supervised object and the identity information corresponding to each supervised object;

[0025] Analysis module: used to analyze the behavior of the supervised object in real time using deep learning methods based on the behavior recognition model, where the behavior of the supervised object includes the detection and classification of abnormal behavior, and obtain the analysis results;

[0026] Judgment module: used to judge the results of the analysis. If an abnormal event is detected, an alarm will be issued and the equipment in the supervision place will be remotely controlled through the system. If no abnormal event is detected, the supervision object will continue to be located and tracked.

[0027] In a third aspect, the present application also provides a surveillance object positioning and tracking device based on video analysis, including:

[0028] Memory for storing computer programs;

[0029] A processor is used to implement the steps of the method for positioning and tracking a supervised object based on video analysis when executing the computer program.

[0030] In a fourth aspect, the present application also provides a readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the above-mentioned method for positioning and tracking a supervised object based on video analysis are implemented.

[0031] The beneficial effects of the present invention are:

[0032] The present invention uses video surveillance analysis to track and locate the position of the supervised object in real time, and displays it in a digital twin model, so as to monitor the dynamics of the supervised object in real time and intelligently analyze the abnormal behavior of the supervised object. The present invention constructs a digital twin model of the supervision site to accurately simulate the space, facilities, equipment and supervision objects of the supervision site. The present invention adopts a deep learning algorithm, based on a convolutional neural network architecture, to train a model to identify the supervised object in the video surveillance of the supervision site, and calculate and update its location information in real time. In combination with a behavior recognition model, a deep learning method is used to perform real-time analysis of the behavior of the supervised object, including detection and classification of abnormal behavior. When an abnormal event is detected, such as abnormal gathering of supervised objects, illegal possession of weapons, climbing, fighting, etc., a corresponding effective alarm notification is generated. The present invention also adopts remote control of equipment: the management personnel of the supervision site can remotely control the equipment in the supervision site through the system, such as access control, and take safety measures in time.

[0033] The present invention targets different missing supervised objects. Supervisors can collect information from different data sources to better track their whereabouts. Video data integration technology can analyze a large number of data sources, including the activity range and behavior patterns of supervised objects, thereby helping supervisors determine possible location information for tracking. The present invention introduces a multi-target tracking algorithm to achieve real-time tracking of multiple targets (supervised objects) in a supervised place, solving the problems of target occlusion and intersection.

[0034] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or be understood by practicing the embodiments of the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description, claims, and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0036] Figure 1 It is a schematic diagram of the flow of the method for positioning and tracking a supervised object based on video analysis described in an embodiment of the present invention;

[0037] Figure 2 It is a schematic diagram of the structure of a supervision object positioning and tracking system based on video analysis described in an embodiment of the present invention;

[0038] Figure 3It is a schematic diagram of the structure of a surveillance object positioning and tracking device based on video analysis described in an embodiment of the present invention.

[0039] In the figure: 701, first acquisition module; 702, second acquisition module; 703, processing module; 7031, segmentation unit; 7032, first extraction unit; 7033, processing unit; 7034, fusion unit; 7035, second extraction unit; 7036, superposition unit; 7037, third extraction unit; 7038, locking unit; 704, analysis module; 705, judgment module; 800, supervision object positioning and tracking device based on video analysis; 801, processor; 802, memory; 803, multimedia component; 804, I / O interface; 805, communication component. DETAILED DESCRIPTION

[0040] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. The components of the embodiments of the present invention generally described and shown in the drawings herein can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the present invention. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0041] It should be noted that similar reference numerals and letters represent similar items in the following drawings, so once an item is defined in one drawing, it does not need to be further defined and explained in the subsequent drawings. At the same time, in the description of the present invention, the terms "first", "second", etc. are only used to distinguish the description and cannot be understood as indicating or implying relative importance.

[0042] Embodiment 1:

[0043] This embodiment provides a method for locating and tracking a supervised object based on video analysis.

[0044] See also Figure 1 , the figure shows that the method includes step S100, step S200, step S300, step S400 and step S500.

[0045] S100. Obtaining a video of a supervised object through a camera, wherein the supervised object movement video includes a supervised object movement segment, a supervised object action segment, and a supervised object activity distribution situation.

[0046] It can be understood that in this step, the supervised object video also includes the real-time positioning of the supervised object, behavior monitoring, trajectory fragments, etc.

[0047] S200. Obtain data information of the supervision site, where the data information includes spatial information, facility information, equipment information and supervision object information of the supervision site, and construct a digital twin model of the supervision site based on the data information of the supervision site.

[0048] It can be understood that in this step, the data information of the regulatory site is acquired to obtain historical regulatory site data information. The historical regulatory site data information can be substituted into the digital twin model of the newly constructed regulatory site for training calculations, and the status of the regulatory site data information can be detected and optimized in real time, providing more accurate data and simulation analysis, which helps to optimize the production and operation processes, improve efficiency, perform virtualization and optimization in a virtual environment, reduce trial and error costs, and realize monitoring and maintenance of remote regulatory sites, reduce manpower and time costs, and improve efficiency.

[0049] S300, processing the mobile video of the supervised object to obtain processed first supervised object information, and based on the convolutional neural network architecture, inputting the first supervised object information into the digital twin model for training, identifying the supervised object in the video surveillance of the supervised place, and obtaining second supervised object information, wherein the second supervised object information includes the location information of each supervised object and the identity information corresponding to each supervised object;

[0050] It can be understood that the step S300 includes S301, S302, S303 and S304, wherein:

[0051] S301, segmenting the moving video of the supervised object, wherein the segmentation process is to extract a segment every five seconds, and each segment includes a moving photo of the supervised object;

[0052] S302, extracting voice data and time-series-based image data from the segments respectively according to the editing software; segmenting the voice data using Bayesian theorem to obtain voice segments;

[0053] It should be noted that the entire speech signal is detected frame by frame, the speech part is compared and judged, and then it is detected according to the Bayesian Information Criterion. By comparing the similarities of adjacent speech signals, the speech situation of the supervised object is judged and segmented to obtain speech segments. The speech segments include the speech of the supervised object, and the speech of the supervised object can be analyzed and judged.

[0054] S303, preprocessing the image data, wherein the preprocessing process includes processing the image data using Gaussian filtering, and optimizing the image data after the noise is weakened according to the edge loss function to obtain image fragments;

[0055] It should be noted that the image can be read after background removal, but the noise in the image may be enhanced. According to the characteristics of the image, we choose Gaussian filtering to filter the image, which can weaken the noise as much as possible and obtain image fragments with high accuracy. The pixel values ​​of each neighborhood are multiplied by the weights of the corresponding positions and then summed. The whole process can also be regarded as a convolution operation between the image and the Gaussian normal distribution.

[0056] S304: The voice segment and the image segment are integrated and summarized to obtain processed first supervision object information.

[0057] By fusing and summarizing voice clips and image clips, we can obtain more complete and accurate first-hand crime information, and lay the foundation for the subsequent positioning and tracking of the supervised objects.

[0058] It should be noted that step S300 also includes S305, S306, S307 and S308, wherein:

[0059] S305, extracting the area range of the movement change of the supervision object target corresponding to the second supervision object information as the target template;

[0060] S306, superimposing the target template and the location of the real-time image, and calculating the corresponding value of the similarity measurement between the two;

[0061] S307, extracting the position corresponding to the maximum similarity in the corresponding values, and then obtaining the position of the supervision object target;

[0062] S308: Lock the location of the supervised object and identify the supervised object information identity based on a preset information database.

[0063] It should be noted that the multi-target tracking algorithm is introduced to achieve real-time tracking of multiple targets (supervisory objects) in the supervision site, solve the problem of target occlusion and intersection, use the information provided by tracking to improve the segmentation effect, match the targets of consecutive frames to track the targets, and the target tracking algorithm uses the global information of the target, such as grayscale, color and texture. Therefore, when the target of the supervision object is not occluded, the tracking accuracy is very high and very stable, the tracking target effect is also good, and the credibility is high.

[0064] S400. According to the behavior recognition model, a deep learning method is used to perform real-time analysis on the behavior of the supervised object, wherein the behavior of the supervised object includes detection and classification of abnormal behavior, and the analysis results are obtained.

[0065] It can be understood that in this step S400, the behavior recognition model includes a classifier, a feature extractor and a decoder, and the behavior recognition model is optimized using a back propagation algorithm and an optimizer, a cross entropy loss function is selected, and the parameters of the behavior recognition model are updated. The training and model tuning are performed through multiple iterations until the behavior recognition model converges and achieves the matching performance.

[0066] It should be noted that choosing the right optimizer and learning rate is also crucial to the performance of the model. The Adam optimizer is often used in deep learning tasks. It has adaptive learning capabilities and momentum characteristics, which can accelerate the convergence of the model. Through different learning rates and combined with learning rate decay strategies, the best model performance can be found to improve the accuracy and stability of the behavior recognition model.

[0067] S500: The analyzed results are judged. If an abnormal event is detected, an alarm is issued, and the internal equipment is remotely controlled through the system. If no abnormal event is detected, the location and tracking of the supervised object continues.

[0068] It can be understood that in this step, the digital security system for supervision places and detention houses is based on functional modules such as video surveillance, access control, perimeter protection, and alarm. In view of the complexity of its environment, the system adopts an integrated security management platform to achieve integrated management and mutual linkage of various systems, effectively reducing security risks and improving the safety protection level of supervision places; and deploying laser radar detectors and laser intrusion detectors at key locations. When illegal targets are found, alarms will be triggered, and sound and light deterrence will be linked. At the same time, the alarm situation will be uploaded to the duty room, and alarm prompts will be given through tracks, voice, icons, and text on the global map. At the same time, the detector's ball camera will accurately track the supervised objects.

[0069] Embodiment 2:

[0070] like Figure 2 As shown, this embodiment provides a supervision object positioning and tracking system based on video analysis, see Figure 2 The system includes a first acquisition module 701, a second acquisition module 702, a processing module 703, an analysis module 704 and a judgment module 705, wherein:

[0071] The first acquisition module 701 is used to acquire a video of a supervised object through a camera, wherein the supervised object movement video includes a supervised object movement segment, a supervised object action segment and a supervised object activity distribution;

[0072] The second acquisition module 702 is used to acquire data information of the supervision site, wherein the data information includes spatial information, facility information, equipment information and supervision object information of the supervision site, and construct a digital twin model of the supervision site based on the data information of the supervision site;

[0073] Processing module 703: used to process the mobile video of the supervised object to obtain the processed first supervised object information; and based on the convolutional neural network architecture, input the first supervised object information into the digital twin model for training, identify the supervised object in the video surveillance of the supervised place, and obtain the second supervised object information, wherein the second supervised object information includes the location information of each supervised object and the identity information corresponding to each supervised object;

[0074] Analysis module 704: used to perform real-time analysis on the behavior of the supervised object using a deep learning method according to the behavior recognition model, wherein the behavior of the supervised object includes detection and classification of abnormal behavior, and obtain the analysis results;

[0075] Judgment module 705: used to judge the results of the analysis. If an abnormal event is detected, an alarm is issued and the equipment in the supervision place is remotely controlled through the system. If no abnormal event is detected, the supervision object is continuously located and tracked.

[0076] Specifically, the processing module 703 includes a segmentation unit 7031, a first extraction unit 7032, a processing unit 7033 and a fusion unit 7034, wherein:

[0077] Segmentation unit 7031: used for segmenting the moving video of the supervised object, wherein the segmentation process is to extract a segment every five seconds, and each segment includes a moving photo of the supervised object;

[0078] The first extraction unit 7032 is used to extract voice data and time-series-based image data from the clips according to the editing software; segment the voice data using the Bayesian theorem to obtain voice clips;

[0079] Processing unit 7033: used for preprocessing the image data, wherein the preprocessing process includes processing the image data by using Gaussian filtering, and optimizing the image data after the noise is weakened according to the edge loss function to obtain the image fragment;

[0080] The fusion unit 7034 is used to fuse and aggregate the voice segment and the image segment to obtain the processed first supervision object information.

[0081] Specifically, the processing module 703 further includes a second extraction unit 7035, a superposition unit 7036, a third extraction unit 7037 and a locking unit 7038, wherein:

[0082] The second extraction unit 7035 is used to extract the area range of the movement change of the supervision object corresponding to the second supervision object information as the target template;

[0083] The superposition unit 7036 is used to superimpose the target template and the position of the real-time image, and calculate the corresponding value of the similarity measurement between the two;

[0084] The third extraction unit 7037 is used to extract the position corresponding to the maximum similarity in the corresponding values, and then obtain the position of the supervision object target;

[0085] Locking unit 7038: used to lock the location of the supervised object target and identify the supervised object information identity based on a preset information database.

[0086] Specifically, the behavior recognition model in the analysis module 704 includes a classifier, a feature extractor and a decoder. The behavior recognition model is optimized using a back-propagation algorithm and an optimizer, a cross-entropy loss function is selected, the parameters of the behavior recognition model are updated, and multiple iterations of training and model tuning are performed until the behavior recognition model converges and achieves the matching performance.

[0087] It should be noted that, regarding the system in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated here.

[0088] Embodiment 3:

[0089] Corresponding to the above method embodiment, this embodiment also provides a supervised object positioning and tracking device based on video analysis. The supervised object positioning and tracking device based on video analysis described below and the supervised object positioning and tracking method based on video analysis described above can refer to each other.

[0090] Figure 3 FIG. 8 is a block diagram of a surveillance object positioning and tracking device 800 based on video analysis according to an exemplary embodiment. Figure 3 As shown, the video analysis-based supervised object positioning and tracking device 800 includes: a processor 801 and a memory 802. The video analysis-based supervised object positioning and tracking device 800 also includes one or more of a multimedia component 803, an I / O interface 804, and a communication component 805.

[0091] The processor 801 is used to control the overall operation of the video analysis-based supervisory object location tracking device 800 to complete all or part of the steps in the video analysis-based supervisory object location tracking method. The memory 802 is used to store various types of data to support the operation of the video analysis-based supervisory object location tracking device 800, which may include, for example, instructions for any application or method operating on the video analysis-based supervisory object location tracking device 800, as well as application-related data, such as contact data, sent and received messages, pictures, audio, video, etc. The memory 802 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, disk or optical disk. The multimedia component 803 may include a screen and an audio component. The screen may be, for example, a touch screen, and the audio component is used to output and / or input audio signals. For example, the audio component may include a microphone for receiving external audio signals. The received audio signal may be further stored in the memory 802 or sent via the communication component 805. The audio component also includes at least one speaker for outputting audio signals. The I / O interface 804 provides an interface between the processor 801 and other interface modules, and the above-mentioned other interface modules can be a keyboard, a mouse or buttons, etc. These buttons can be virtual buttons or physical buttons. The communication component 805 is used for wired or wireless communication between the video analysis-based supervision object positioning and tracking device 800 and other devices. Wireless communication, such as Wi-Fi, Bluetooth, Near Field Communication (NFC), 2G, 3G or 4G, or a combination of one or more of them, so the corresponding communication component 805 can include: a Wi-Fi module, a Bluetooth module or an NFC module.

[0092] In an exemplary embodiment, the video analysis-based supervised object positioning and tracking device 800 can be implemented by one or more application specific integrated circuits (Application Specific Integrated Circuit, referred to as ASIC), digital signal processors (Digital Signal Processor, referred to as DSP), digital signal processing devices (Digital Signal Processing Device, referred to as DSPD), programmable logic devices (Programmable Logic Device, referred to as PLD), field programmable gate arrays (Field Programmable Gate Array, referred to as FPGA), controllers, microcontrollers, microprocessors or other electronic components to execute the above-mentioned video analysis-based supervised object positioning and tracking method.

[0093] In another exemplary embodiment, a computer-readable storage medium including program instructions is also provided, and when the program instructions are executed by a processor, the steps of the above-mentioned method for positioning and tracking a supervised object based on video analysis are implemented. For example, the computer-readable storage medium can be the above-mentioned memory 802 including program instructions, and the above-mentioned program instructions can be executed by the processor 801 of the device 800 for positioning and tracking a supervised object based on video analysis to complete the above-mentioned method for positioning and tracking a supervised object based on video analysis.

[0094] Embodiment 4:

[0095] Corresponding to the above method embodiment, a readable storage medium is also provided in this embodiment. The readable storage medium described below and the method for positioning and tracking a supervised object based on video analysis described above can refer to each other.

[0096] A computer program is stored on the readable storage medium, and when the computer program is executed by the processor, the steps of the method for positioning and tracking a supervised object based on video analysis in the above-mentioned method embodiment are implemented.

[0097] The readable storage medium may specifically be a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, or other readable storage medium that can store program codes.

[0098] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

[0099] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A method for locating and tracking a supervised object based on video analysis, characterized in that: include: Obtaining videos of supervised objects through cameras, wherein the supervised object movement videos include supervised object movement clips, supervised object action clips, and supervised object activity distribution; Obtaining data information of the supervision site, where the data information includes spatial information, facility information, equipment information and supervision object information of the supervision site, and building a digital twin model of the supervision site based on the data information of the supervision site; Processing the mobile video of the supervised object to obtain processed first supervised object information; and based on the convolutional neural network architecture, inputting the first supervised object information into the digital twin model for training, identifying the supervised object in the video surveillance of the supervised place, and obtaining second supervised object information, wherein the second supervised object information includes the location information of each supervised object and the identity information corresponding to each supervised object; Based on the behavior recognition model, the deep learning method is used to analyze the behavior of the supervised object in real time, where the behavior of the supervised object includes the detection and classification of abnormal behavior, and the analysis results are obtained; The results of the analysis are judged. If an abnormal event is detected, an alarm is issued and the equipment in the supervision place is remotely controlled through the system. If no abnormal event is detected, the supervision object will continue to be located and tracked.

2. The method for locating and tracking a supervised object based on video analysis according to claim 1, characterized in that: The mobile video of the supervised object is processed to obtain the processed first supervised object information, which includes: Perform video segmentation on the moving video of the supervised object, wherein the segmentation process is to extract a segment every five seconds, and each segment includes a photo of the supervised object's activities; Extract voice data and time-series-based image data from the clips using editing software; segment the voice data using Bayesian theorem to obtain voice clips; Preprocessing the image data, wherein the preprocessing process includes processing the image data using Gaussian filtering, and optimizing the image data after the noise is weakened according to the edge loss function to obtain image fragments; The voice segment and the image segment are fused and aggregated to obtain processed first supervision object information.

3. The method for positioning and tracking a supervised object based on video analysis according to claim 1, characterized in that: The second supervision object information is obtained, including: Extracting the area range of the motion change of the supervision object target corresponding to the second supervision object information as the target template; Superimpose the target template and the real-time image location, and calculate the corresponding value of the similarity measure between the two; The position corresponding to the maximum similarity in the corresponding values ​​is extracted, and then the position of the target of the supervision object is obtained; The location of the supervised object target is locked, and the supervised object information identity is identified based on the preset information database.

4. The method for locating and tracking a supervised object based on video analysis according to claim 1, characterized in that: The behavior recognition model includes a classifier, a feature extractor and a decoder. The behavior recognition model is optimized using a back-propagation algorithm and an optimizer, a cross-entropy loss function is selected, the parameters of the behavior recognition model are updated, and multiple iterations of training and model tuning are performed until the behavior recognition model converges and achieves the matching performance.

5. The supervision object positioning and tracking system based on video analysis is characterized by: include: The first acquisition module is used to acquire the video of the supervised object through the camera, wherein the supervised object movement video includes the supervised object movement segment, the supervised object action segment and the supervised object activity distribution; The second acquisition module is used to acquire data information of the supervision site, where the data information includes spatial information, facility information, equipment information and supervision object information of the supervision site, and to build a digital twin model of the supervision site based on the data information of the supervision site; Processing module: used to process the mobile video of the supervised object to obtain the processed first supervised object information; and based on the convolutional neural network architecture, input the first supervised object information into the digital twin model for training, identify the supervised object in the video surveillance of the supervised place, and obtain the second supervised object information, wherein the second supervised object information includes the location information of each supervised object and the identity information corresponding to each supervised object; Analysis module: used to analyze the behavior of the supervised object in real time using deep learning methods based on the behavior recognition model, where the behavior of the supervised object includes the detection and classification of abnormal behavior, and obtain the analysis results; Judgment module: used to judge the results of the analysis. If an abnormal event is detected, an alarm will be issued and the equipment in the supervision place will be remotely controlled through the system. If no abnormal event is detected, the supervision object will continue to be located and tracked.

6. The video analysis-based supervision object positioning and tracking system according to claim 5 is characterized in that: The processing module includes: Segmentation unit: used to segment the moving video of the supervised object, wherein the segmentation process is to extract a segment every five seconds, and each segment includes a moving photo of the supervised object; The first extraction unit is used to extract voice data and image data based on time series from the clips according to the editing software; the voice data is segmented using the Bayesian theorem to obtain voice clips; Processing unit: used for preprocessing the image data, wherein the preprocessing process includes processing the image data by using Gaussian filtering, and optimizing the image data after the noise is weakened according to the edge loss function to obtain the image fragment; Fusion unit: used to fuse and aggregate the voice segment and the image segment to obtain the processed first supervision object information.

7. The video analysis-based supervision object positioning and tracking system according to claim 5, characterized in that: The processing module further comprises: A second extraction unit: used for extracting a region range of motion change of a supervision object target corresponding to the second supervision object information as a target template; Overlay unit: used to overlay the target template with the location of the real-time image and calculate the corresponding value of the similarity measure between the two; The third extraction unit is used to extract the position corresponding to the maximum similarity in the corresponding values, so as to obtain the position of the supervision object target; Locking unit: used to lock the location of the supervised object and identify the supervised object's information identity based on the preset information database.

8. The video analysis-based supervision object positioning and tracking system according to claim 5, characterized in that: The behavior recognition model in the analysis module includes a classifier, a feature extractor and a decoder. The behavior recognition model is optimized using a back-propagation algorithm and an optimizer, a cross-entropy loss function is selected, the parameters of the behavior recognition model are updated, and multiple iterations of training and model tuning are performed until the behavior recognition model converges and achieves the matching performance.

9. A surveillance object positioning and tracking device based on video analysis, characterized in that: include: Memory for storing computer programs; A processor, configured to implement the method for positioning and tracking a supervised object based on video analysis as described in any one of claims 1 to 4 when executing the computer program.

10. A readable storage medium, characterized in that: The readable storage medium stores a computer program, and when the computer program is executed by a processor, the method for positioning and tracking a supervised object based on video analysis as described in any one of claims 1 to 4 is implemented.

Images