Examination room and security room behavior analysis patrol method and system using intelligent agent

Abstract

The invention relates to the technical field of data management, and discloses an examination room and security room behavior analysis patrol method and system using an intelligent agent, which can focus on the maintenance and management of resources required by an examination in the examination room preparation stage to form high-quality data resources which can be used by a service system. Through communication with a data interface of a comprehensive management platform, effectiveness and accuracy of data are ensured, a mapping relation is formed between identity coding information and examination places in an examination plan by using the received examination plan, so that functions during an examination focus on process control and supervision, false alarm reduction and identification analysis accuracy improvement, and a reexamination instruction is used to improve the accuracy of examination. And the extracted behavior events are rechecked, the conclusion of each behavior event is fed back to the conclusion display end, effective decision support is provided for examination affairs, the intelligent patrol invigilation client is matched, post-event tracing is changed into in-event real-time disposal, and the timeliness of patrol rechecking work is improved.

Description

Technical Field

[0001] This invention relates to a method and system for analyzing and inspecting examination rooms and confidential rooms using intelligent agents, belonging to the field of data management technology. Background Technology

[0002] With the popularization of computer vision and IoT technologies, video surveillance equipment has been widely deployed in various scenarios such as national education examinations and professional qualification examinations to achieve online inspections, forming a new invigilation model that combines human and technological security measures. Current mainstream AI-powered invigilation systems have achieved intelligent identification of basic violations, significantly improving inspection efficiency and providing effective guarantees for the fairness of examinations. However, in practical applications, there is still room for optimization in the refined supervision of complex scenarios. Regarding equipment management and scheduling, while the camera equipment in examination rooms and secure rooms has basic identification capabilities, the dynamic association mechanism between standardized identification codes and examination plans is still not perfect. When an examination involves multiple areas, multiple examination rooms, and dedicated secure rooms, invigilators need to manually match target equipment among numerous monitoring links, and the accuracy and timeliness of equipment scheduling need to be improved.

[0003] Therefore, it is necessary to address the shortcomings of traditional online inspection and review methods, which involve a large workload and result in low timeliness of review conclusions. A new inspection method and system that utilizes intelligent agents to analyze the behavior of examiners in examination rooms and secure rooms should be proposed. Summary of the Invention

[0004] This invention provides a method and system for analyzing and inspecting examination rooms and confidential rooms using intelligent agents, which can solve the problem of low timeliness of review conclusions due to the large workload of online inspection and verification.

[0005] This invention provides a method for monitoring examination rooms and secure rooms using intelligent agents through behavioral analysis, comprising: Receive network access requests from video capture devices; Based on the network access request of the acquisition camera device, obtain the identity code information of the acquisition camera device; Using the received examination plan, a mapping relationship is established between the identity code information and the examination room sessions in the examination plan; Select the identification code information of the camera device used to monitor the examination room during the exam; Using identity code information, real-time video of the examination room being monitored during the exam can be obtained; Based on the recognition instructions, behavioral events are extracted from real-time video. Return to the step of selecting the identity code information of the video capture device for monitoring the examination room during the examination, until all identity code information has been selected; The review order is used to review the extracted behavioral events; The conclusion for each behavioral event is fed back to the conclusion display terminal.

[0006] This invention provides a monitoring system for examination rooms and secure rooms that utilizes intelligent agents to analyze and patrol behavior, comprising: The server is used to execute the above-mentioned method of using intelligent agents to analyze and inspect examination rooms and confidential rooms. The camera device is connected to the server for video capture.

[0007] This invention discloses a method and system for analyzing and monitoring examination room and secure room behavior using intelligent agents. By receiving network access requests from video acquisition devices and obtaining their identification codes, the system focuses on maintaining and managing resources needed for the examination during the examination preparation phase, generating high-quality data resources usable by business systems. Communication with the integrated management platform's data interface ensures data validity and accuracy. Using the received examination plan, the system maps the identification codes to the examination sessions, allowing for process control and supervision during the examination period. It guides behavior at key moments, provides timely feedback, integration, statistics, and display of various data generated during execution, and enables timely decision-making regarding risks and deviations. Based on identification commands, the system extracts behavioral events from real-time video, obtaining real-time video of the examination room during the examination. Combined with an intelligent monitoring client, it facilitates intelligent monitoring and supervision arrangements, transforming post-event traceability into real-time, in-process handling. The machine vision-based deep learning behavior analysis algorithm is trained using an open-source deep learning framework and continuously iterates and updates to reduce false alarms and improve the accuracy of recognition and analysis. It utilizes review instructions to review extracted behavioral events and provides feedback on the conclusions of each event to the display terminal. Taking into account the principle of separating business processes from component implementation, it uses process management, strategy management, and interface integration technologies to dynamically define system behavior to achieve system functions. The intelligent inspection system platform and system, along with alarm information reported level by level to the examination platform, provide effective decision support for examination administration. Combined with the intelligent inspection and invigilation client, it transforms post-event traceability into real-time handling during the event, improving the timeliness of inspection and review work. Attached Figure Description

[0008] Figure 1 This is a flowchart illustrating a method for analyzing and inspecting examination rooms and secure rooms using intelligent agents, according to an embodiment of the present invention. Figure 2 This is a system connection diagram of an examination room and confidential room behavior analysis and inspection system utilizing intelligent agents, according to one embodiment of the present invention. Icon labels: 100 - Server; 200 - Camera / video capture equipment. Detailed Implementation

[0009] To make the objectives, technical solutions, and advantages of the present invention clearer, the technical solutions of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0010] like Figure 1 As shown, the present invention provides a method for monitoring examination rooms and confidential rooms using intelligent agents through behavioral analysis, comprising: S100 receives network access requests from video acquisition cameras.

[0011] S200 obtains the identity code information of the acquisition camera device based on the network access request of the acquisition camera device.

[0012] S300 uses the received examination plan to map the identity code information to the examination room sessions in the examination plan.

[0013] S400, select the identification code information of the video capture device that monitors the examination room during the examination.

[0014] The S500 uses identity coding information to obtain real-time video of the examination room during the exam.

[0015] The S600 extracts behavioral events from real-time video based on recognition commands.

[0016] S700, return to the step of selecting the identity code information of the video capture device for monitoring the examination room during the examination, until all identity code information has been selected.

[0017] S800 uses review instructions to review extracted behavioral events.

[0018] S900 feeds back the conclusion of each behavioral event to the conclusion display terminal.

[0019] Specifically, based on artificial intelligence and deep learning technologies, combined with the online inspection system's video transmission architecture, the complete architecture design includes a real-time intelligent inspection platform for the upper level and a real-time intelligent inspection platform for the confidential room, a multimodal behavior analysis super brain and behavior analysis review for the school-level examination room or confidential room, intelligent video inspectors, and an access gateway. The behavior analysis super brain actively pushes suspicious and abnormal behavior results to the video inspectors, greatly reducing the workload of manually monitoring the examination videos.

[0020] The real-time intelligent inspection and confidential room real-time intelligent inspection platform introduces intelligent video inspectors into the intelligent inspection system architecture, transforming post-event review into in-event handling, so that the results of intelligent behavior analysis can provide real-time basis and support for the examination room emergency command and examination affairs decision-making system.

[0021] The real-time patrol and secure room real-time intelligent patrol platform mainly comprises three functional modules: an intelligent video proctoring management module, a behavior analysis management module, and an intelligent video proctor client; and two communication modules: a unified access platform and a school-level intelligent gateway. The intelligent video proctoring management module utilizes the existing online patrol system and simultaneously acquires basic data and examination data provided by the comprehensive management platform, providing fundamental data and functional support for the entire platform. The behavior analysis management module manages and schedules the installed and deployed super-brain devices, cloud analysis units, and other computing resources, ensuring their rational and efficient use. The video proctor client is used by video proctors. Simultaneously, through message queues, SDK APIs, and other interfaces, various data are pushed to the comprehensive management platform, providing support for higher-level decision-making and consultation. The multimodal behavior analysis super-brain interfaces with the higher-level unified access platform through the deployed school-level intelligent access gateway device via a physical model, enabling the reporting of abnormal behavior data and the issuance of some control commands.

[0022] This application relates to a method for analyzing and monitoring examination room and secure room behavior using intelligent agents. By receiving network access requests from video recording devices, the method obtains the device's identification code information. During the examination preparation phase, the method focuses on maintaining and managing resources needed for the examination, creating high-quality data resources usable by business systems. Communication with the integrated management platform's data interface ensures data validity and accuracy. Using the received examination plan, the method maps the identification code information to the examination room sessions, allowing the method to focus on process control and supervision during the examination. It guides behavior at key time points, provides timely feedback, integration, statistics, and display of various data generated during execution, and enables timely decision-making regarding risks and deviations. Using the identification code information and based on recognition commands, the method extracts behavioral events from real-time video, obtaining real-time video of the examination room during the examination. Combined with an intelligent monitoring and invigilation client, it facilitates intelligent monitoring and invigilation arrangements, transforming post-event traceability into real-time in-event handling. The machine vision-based deep learning behavior analysis algorithm is trained using an open-source deep learning framework and continuously iterates and updates to reduce false alarms and improve the accuracy of recognition and analysis. It utilizes review instructions to review extracted behavioral events and provides feedback on the conclusions of each event to the display terminal. Taking into account the principle of separating business processes from component implementation, it uses process management, strategy management, and interface integration technologies to dynamically define system behavior to achieve system functions. The intelligent inspection system platform and system, along with alarm information reported level by level to the examination platform, provide effective decision support for examination administration. Combined with the intelligent inspection and invigilation client, it transforms post-event traceability into real-time handling during the event, improving the timeliness of inspection and review work.

[0023] In one embodiment of this application, S100 includes: S110 receives the examination room or secure room number and the monitoring equipment position number.

[0024] S120 establishes a mapping relationship between the examination room or secure room number and the monitoring equipment position number.

[0025] S130 receives network access requests from video acquisition cameras.

[0026] S140, based on the network access request of the acquisition camera device, determines the video allocation level of the acquisition camera device.

[0027] S150 connects the data stream of the acquisition camera to the video platform according to the video allocation level of the acquisition camera.

[0028] S160, search for the exam room or secure room number on the video platform.

[0029] S170: Using the search results from the video platform, determine whether the level of the acquisition camera connected to the video platform matches the level of the monitoring equipment's location number.

[0030] S180: If the level of the acquisition camera connected to the video platform matches the level of the monitoring equipment's location number, then feedback will be sent indicating that the acquisition camera has successfully joined the network.

[0031] S190 If the level of the acquisition camera connected to the video platform does not match the level of the monitoring equipment's location number, then feedback will be sent indicating an abnormal network access for the acquisition camera.

[0032] Understandably, multiple monitoring camera positions are set up in the examination room or secure room, and the examination room or secure room number can be associated with the monitoring camera position number in the video platform. The acquisition camera device sending the network access request supports HDMI 4K signal input and encoding. When the acquisition camera device connected to the video platform meets the requirements for HDMI input and encoding, a successful network access message is sent. The acquisition camera device sending the network access request will act as a monitoring device in the examination room or secure room, monitoring the behavior of personnel in the examination room or secure room, including physical actions and verbal behavior.

[0033] If the level of the acquisition camera connected to the video platform does not match the level of the monitoring equipment's location number, then the acquisition camera needs to be repaired by equipment maintenance personnel.

[0034] In this embodiment, the examination supervision platform design, which maps the examination room or secure room number to the monitoring equipment position number, and the intelligent inspection function of the examination site's secure room are designed independently and loosely coupled to avoid the paralysis of the examination site's business functions due to network failures on the platform side or the examination site side. Searching for the examination room or secure room number on the video platform avoids the problem of mutual omissions between the examination room / secure room and the video platform, ensuring network stability. Data distribution is completed after the examination room arrangement, depending on the network conditions. Data reporting can be cached in case of network downtime and reliably uploaded after the network is restored.

[0035] In one embodiment of this application, S200 includes: S210 invokes a network access request from a successfully connected video capture device.

[0036] S220, analyzes the network access request of the camera device.

[0037] S230, obtain the identification code information of the camera device.

[0038] S240 sends the identification code information to the video inspection system.

[0039] S250, based on the examination site SIP routing gateway, locates the monitoring device's location number.

[0040] S260, determine whether the identity code information matches the monitoring device's location number.

[0041] S270 If the identity code information matches the monitoring equipment position number, the configuration of the video capture equipment in the examination room or confidential room is completed.

[0042] S280 If the identity code information does not match the monitoring equipment location number, feedback will be sent indicating that the configuration of the video capture equipment in the examination room or confidential room has failed.

[0043] S290, return to the network access request of a successfully accessing video capture device, until all network access requests have been selected and completed.

[0044] Specifically, a SIP routing gateway is a network device function, typically used in routers, firewalls, and similar devices. It can inspect and modify SIP (Session Initiation Protocol) signaling and media streams to resolve communication issues in Network Address Translation (NAT) environments. When a NAT device translates a private IP address to a public IP address, the private IP address may be embedded in the SIP packet. The SIP routing gateway parses these SIP headers, replacing the embedded private IP with a public IP address outside the network address, ensuring the receiver knows the correct return address.

[0045] In this embodiment, the multimodal supercomputer for behavioral analysis in the examination room and secure room employs a distributed design for key functional modules, a distributed decoding design, and a distributed computing power design, with each module serving as a backup. This ensures long-term stable system operation, strong fault tolerance, and system recovery capabilities, preventing system collapse due to a single point of failure. A unified video platform interfaces with the university-level intelligent access gateway device, enabling data communication between the higher-level and university-level systems through a physical model interface. It can monitor the real-time dynamics of the multimodal behavioral analysis supercomputer, receive various suspicious cheating information and push and distribute it to various functional modules, and receive and issue various commands.

[0046] The main security functions of the intelligent patrol platform include: using the identity code information of the collected camera equipment, device identity authentication of the examination site SIP routing gateway, security status detection, access control, security situation analysis, user identity authentication, security function management, remote management security, audit and log security, and security controllability.

[0047] In one embodiment of this application, S300 includes: S310, receiving the examination plan.

[0048] S320 obtains the handheld client information of the invigilator, the conclusion display terminal information of the inspection administrator, and the number of the examination room or secure room to be used in the examination plan.

[0049] S330: Based on the number of the examination room or secure room to be used, obtain the identity code information corresponding to the number of the examination room or secure room.

[0050] S340 uses the level of the invigilator's handheld client information to send the URL in the identity code information to the invigilator's handheld client.

[0051] The S350 utilizes the information from the inspection administrator's conclusion display terminal to connect the inspection administrator's conclusion display terminal to the video inspection system.

[0052] Understandably, the handheld client information of invigilators is mainly operated by video invigilators at all levels. Through the provided functions such as video inspection, examination guidance, on-duty checks, and status checks, they can conveniently, efficiently, and accurately fulfill the various responsibilities required in the examination manual. The intelligent video invigilator client introduced into the intelligent inspection system architecture, in conjunction with the invigilation settings of video invigilators, is to transform post-event review into in-event review.

[0053] A URL, or Uniform Resource Locator, is the standard way to locate and access various resources on the internet. It consists of a series of characters, including protocols such as HTTP and HTTPS, domain names, paths, and query parameters. These elements together form a complete URL address, enabling computers and users to accurately find and access specific information or services. Every resource on the internet has a unique URL indicating its location and access method. Utilizing the proctor's handheld client information level, the URL from the identity encoding information is sent to the proctor's handheld client. The proctor can then connect to the URL through the handheld client to monitor abnormal behavior in the examination room or secure room in real time.

[0054] In this embodiment, the invigilator's handheld client and the inspection administrator's conclusion display terminal are connected to the existing education inspection system to ensure resource sharing. The system should adopt a modular design, with easily expandable system scale and functions. The system's supporting software should have the ability to be upgraded and integrated with other systems. SIP-based authentication and networking are used for seamless integration with the education inspection system.

[0055] In one embodiment of this application, S500 includes: The S510 receives real-time video of the examination room during the exam, based on the identity code information.

[0056] S520 receives behavior recognition tasks.

[0057] The S530 captures image frames from real-time video.

[0058] S540 arranges image frames sequentially according to time sequence.

[0059] S550 compares the poses of people in adjacent image frames.

[0060] S560 marks the time intervals in which the posture of people changes in adjacent image frames.

[0061] The S570 provides feedback on real-time video clips with time interval labels.

[0062] Understandably, the behavior analysis management module mainly consists of analysis task management and basic data management. Relying on big data technology and artificial intelligence deep learning networks, it allocates tasks to computing resources such as the super brain and submits test results, assisting video proctors in efficiently fulfilling their proctoring duties, reducing their workload, and also providing an effective means for consultation and judgment by higher-level departments.

[0063] The behavior analysis management module manages and schedules the installed and deployed super-brain devices, cloud analysis units, and other computing resources to ensure their rational and efficient use. The intelligent video proctor client is used by video proctors and provides standardized, accurate, and easy-to-use operating functions according to the responsibilities outlined in the examination manual. It also offers various guidance options to assist proctors in completing their tasks promptly, efficiently, and completely. Simultaneously, through message queues, SDK APIs, and other interfaces, it pushes various data to the integrated management platform, supporting decision-making and consultation with higher-level departments. The university-level multimodal behavior analysis super-brain interfaces with the higher-level unified access platform through deployed university-level intelligent access gateway devices via a physical model, enabling the reporting of abnormal behavior data and the issuance of some control commands.

[0064] In one embodiment of this application, S600 includes: S611, Select a video segment of a real-time video with a time interval label.

[0065] S612 incorporates the selected video clips into the cloud detection unit of the video inspection system.

[0066] S613 determines whether the selected video segment is a behavioral event based on the output features of the cloud detection unit.

[0067] S614, if the selected video segment is an action event, then the selected video segment is added to the pending reporting folder, and the process of selecting a real-time video segment with a time interval label is repeated until all video segments have been selected.

[0068] S615, if the selected video segment is not an action event, then return to the step of selecting a real-time video segment with a time interval label, until all video segments have been selected.

[0069] Understandably, the dual-model approach employs two models: one for the exam and one for the post-exam period. The larger model enables intelligent review with low workload, while the smaller model used during the exam utilizes cloud detection units, offering better real-time performance and faster, more comprehensive behavior analysis. Before manual review, intelligent review can be configured based on the larger model, ensuring accurate behavior analysis and reducing the workload of entirely manual review.

[0070] In this embodiment, the video acquisition device supports speech recognition. It combines video-based behavior recognition with a speech deep learning model that utilizes the examination room audio pickup to perform compliance analysis of speech in accordance with the examination manual, and supports the analysis of abnormal speech by invigilators.

[0071] In one embodiment of this application, S600 further includes: S621 receives feedback information from the invigilator's handheld client.

[0072] S622 will include suspicious video clips reported by the handheld client into the pending reporting folder.

[0073] Understandably, utilizing suspicious video clips fed back by invigilators' handheld devices enhances the data sources for abnormal behavior analysis and improves the information coverage and diversity of abnormal behavior inspections in examination rooms.

[0074] In one embodiment of this application, S800 includes: S811 receives motion posture images of abnormal behavior based on re-inspection commands.

[0075] S812 is a comprehensive image algorithm that incorporates motion and posture images of abnormal behavior into the video inspection system.

[0076] S813, obtain the characteristics for determining abnormal behavior.

[0077] S814, select a video clip from the folder to be reported.

[0078] S815 incorporates selected video clips into a comprehensive image algorithm.

[0079] S816 uses the characteristics of abnormal behavior to obtain the judgment conclusion of the selected video segment.

[0080] S817, return to the selected video segment in the folder to be reported, until all video segments have been selected.

[0081] Understandably, the analysis of exam room and confidential room behavior is comprehensive, including the analysis of invigilator compliance and the analysis of suspicious behavior of candidates in the exam room.

[0082] The system employs advanced and mature technologies to construct its various subsystems, forming a stable and reliable system. A single algorithm simultaneously covers the analysis and recognition of examinee behavior in the examination room and the analysis of teacher invigilation compliance, enabling it to operate efficiently, securely, and smoothly. The architecture design supports real-time video stream analysis and recognition, as well as historical video stream analysis and recognition.

[0083] In one embodiment of this application, S800 further includes: S821 receives the judgment conclusion for each video segment in the folder to be reported.

[0084] S822, call the number of the examination room or confidential room.

[0085] S823, select a judgment conclusion for a video segment.

[0086] S824, based on the identification code information of the camera device capturing the selected video clip, finds the number of the examination room or confidential room.

[0087] S825 establishes a one-to-one mapping relationship between the examination room or secure room number and the video clip.

[0088] S826, return to the determination result of selecting a video segment, until all determination results have been selected.

[0089] S827, generate statistical event data to be fed back.

[0090] Understandably, large models reduce the workload of intelligent review. Two models are used during and after the exam: a small model from the cloud detection unit during the exam, which offers good real-time performance and fast and comprehensive behavior analysis. After the exam, a large model based on Transformer can be configured for intelligent review, ensuring accurate behavior analysis and thus reducing the workload of entirely manual review.

[0091] By leveraging technologies such as computer vision, image processing, pattern recognition, and deep learning, and through the access and analysis of real-time video and audio streams, this system enables the detection of suspected abnormal behavior by examinees during the exam. The system then assesses the detection results, ultimately confirming, processing, statistically analyzing, and displaying information related to cheating, thus constructing a high-quality, information-based intelligent proctoring system. The system also supports video-based analysis of examinee behavior, detecting abnormal behavior during the exam and issuing alerts for such behavior.

[0092] The unified access platform interfaces with the school-level intelligent access gateway device, enabling data communication between the higher-level and school-level systems through a physical model interface. It can monitor the real-time dynamics of the multimodal behavior analysis supercomputer, receive various suspicious information and push and distribute it to various functional modules, and receive and issue various commands.

[0093] like Figure 2 As shown, the present invention provides an examination room and confidential room behavior analysis and inspection system using intelligent agents, comprising: Server 100 is used to execute the above-mentioned method of using intelligent agents to analyze and inspect examination rooms and confidential rooms.

[0094] The video acquisition camera 200 is communicatively connected to the server 100.

[0095] Specifically, server 100 can connect to the invigilator's handheld client and the inspection administrator's conclusion display terminal via a network connection channel.

[0096] This embodiment relates to an examination room and secure room behavior analysis and inspection system utilizing intelligent agents. The server 100 receives network access requests from the acquisition camera 200 and obtains the identity code information of the acquisition camera 200. During the examination preparation phase, the system focuses on maintaining and managing the resources required for the examination, forming high-quality data resources usable by the business system. Communication with the integrated management platform's data interface ensures the validity and accuracy of the data. Using the received examination plan, the server maps the identity code information to the examination sessions in the plan, allowing the system to focus on process control and supervision during the examination. The server 100 guides behavior at key time points, promptly feedbacks, integrates, statistically analyzes, and displays various data generated during execution, and makes timely decisions regarding risks and deviations. Using the identity code information and based on recognition commands, the system extracts behavioral events from real-time video, obtaining real-time video of the examination room during the examination. The server 100, in conjunction with the intelligent inspection and invigilation client, makes intelligent inspection and invigilation arrangements, transforming post-event traceability into real-time in-event handling. The machine vision-based deep learning behavior analysis algorithm is trained using an open-source deep learning framework and continuously iterates and updates to reduce false alarms and improve the accuracy of recognition and analysis. It utilizes review instructions to review extracted behavioral events and provides feedback on the conclusions of each event to the display terminal. Taking into account the principle of separating business processes from component implementation, it uses process management, strategy management, and interface integration technologies to dynamically define system behavior to achieve system functions. The intelligent inspection system platform and system, along with alarm information reported level by level to the examination platform, provide effective decision support for examination administration. Combined with the intelligent inspection and invigilation client, it transforms post-event traceability into real-time handling during the event, improving the timeliness of inspection and review work.

[0097] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0098] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for monitoring examination rooms and secure rooms using intelligent agents based on behavioral analysis, characterized in that, include: Receive network access requests from video capture devices; Based on the network access request of the acquisition camera device, obtain the identity code information of the acquisition camera device; Using the received examination plan, a mapping relationship is established between the identity code information and the examination room sessions in the examination plan; Select the identification code information of the camera device used to monitor the examination room during the exam; Using identity code information, real-time video of the examination room being monitored during the exam can be obtained; Based on the recognition instructions, behavioral events are extracted from real-time video. Return to the step of selecting the identity code information of the video capture device for monitoring the examination room during the examination, until all identity code information has been selected; The review order is used to review the extracted behavioral events; The conclusion for each behavioral event is fed back to the conclusion display terminal.

2. The method for analyzing and patrolling examination rooms and confidential rooms using intelligent agents according to claim 1, characterized in that, The process of receiving network access requests from video acquisition devices includes: Receive the examination room or secure room number and the monitoring equipment location number; Establish a mapping relationship between the examination room or secure room number and the monitoring equipment position number; Receive network access requests from video capture devices; Based on the network access request of the acquisition camera equipment, determine the video allocation level of the acquisition camera equipment; Based on the video allocation level of the acquisition camera equipment, the data stream of the acquisition camera equipment is connected to the video platform; Search for the exam room or secure room number on the video platform; Using the search results from the video platform, determine whether the level of the acquisition camera connected to the video platform matches the level of the monitoring equipment's location number; If the level of the acquisition camera connected to the video platform matches the level of the monitoring equipment's location number, then a message indicating that the acquisition camera has successfully joined the network will be sent. If the level of the acquisition camera connected to the video platform does not match the level of the monitoring equipment's location number, a network access error message for the acquisition camera will be reported.

3. The method for monitoring examination rooms and confidential rooms using intelligent agents based on behavioral analysis according to claim 2, characterized in that, The process of obtaining the identity code information of the acquisition camera device based on its network access request includes: Invoke a network access request from a successfully connected video capture device; Analyze the network access request from the video capture device; Obtain the identification code information of the video capture device; Send the identification code information to the video inspection system; Based on the SIP routing gateway at the test site, locate the monitoring device's mounting point number; Determine whether the identity code information matches the monitoring device's location number; If the identity code information matches the monitoring equipment position number, the configuration of the video capture equipment in the examination room or confidential room is completed. If the identity code information does not match the monitoring equipment position number, a message will be sent indicating that the configuration of the video capture equipment in the examination room or confidential room has failed. Return to the network access request of a successfully connected video capture device, until all network access requests have been completed.

4. The method for monitoring examination rooms and confidential rooms using intelligent agents based on behavioral analysis according to claim 3, characterized in that, The process of mapping identity coding information to examination sessions in the received examination plan includes: Receive the exam plan; Obtain the handheld client information of the invigilators, the conclusion display terminal information of the inspection administrator, and the number of the examination room or secure room to be used in the examination plan; Based on the number of the examination room or secure room to be used, obtain the identity code information corresponding to the number of the examination room or secure room; By utilizing the level of the invigilator's handheld client information, the URL in the identity code information is sent to the invigilator's handheld client; By utilizing the conclusion display information of the inspection administrator, the conclusion display terminal of the inspection administrator is connected to the video inspection system.

5. The method for monitoring examination rooms and confidential rooms using intelligent agents based on behavioral analysis according to claim 4, characterized in that, The method of obtaining real-time video of the examination room during the examination using identity coding information includes: Based on the identity code information, receive real-time video of the examination room during the exam; Receive behavior recognition task; Capture image frames from real-time video; The image frames are arranged in chronological order. Compare the poses of people in adjacent image frames; Mark the time intervals of changes in human posture in adjacent image frames; Feedback includes real-time video clips with time interval labels.

6. The method for monitoring examination rooms and confidential rooms using intelligent agents based on behavioral analysis according to claim 5, characterized in that, The extraction of behavioral events from real-time video based on recognition instructions includes: Select a video segment from a real-time video with a time interval label; The selected video clips are incorporated into the cloud detection unit of the video inspection system; Based on the output features of the cloud detection unit, it is determined whether the selected video segment is a behavioral event; If the selected video segment is an action event, the selected video segment will be added to the pending reporting folder, and the process of selecting a real-time video segment with a time interval label will be repeated until all video segments have been selected. If the selected video segment is not an action event, then return to the previous step of selecting a video segment of a real-time video with a time interval label, until all video segments have been selected.

7. The method for monitoring examination rooms and confidential rooms using intelligent agents based on behavioral analysis according to claim 6, characterized in that, The extraction of behavioral events from real-time video based on recognition instructions also includes: Receive feedback information from the invigilator's handheld client; Suspicious video clips reported by the handheld client will be added to the pending reporting folder.

8. The method for monitoring examination rooms and confidential rooms using intelligent agents based on behavioral analysis according to claim 7, characterized in that, The process of reviewing extracted behavioral events using review instructions includes: Based on the re-inspection command, receive motion posture images of abnormal behaviors; Incorporate motion and posture images of abnormal behavior into the comprehensive image algorithm of the video inspection system; Obtain the identifying characteristics of abnormal behavior; Select a video clip from the folder to be reported; The selected video clips are incorporated into the comprehensive image algorithm; By utilizing the characteristics of abnormal behavior, a judgment conclusion can be obtained for the selected video segment; Return to the folder where you selected a video segment to be reported, and continue until all video segments have been selected.

9. The method for monitoring examination rooms and confidential rooms using intelligent agents based on behavioral analysis according to claim 8, characterized in that, The method of reviewing extracted behavioral events using review instructions also includes: Receive the judgment results for each video segment in the folder to be reported; Call the examination room or secure room number; Select a video segment and draw a conclusion; Based on the identification code information of the camera equipment used to capture the selected video clips, the number of the examination room or confidential room can be found. Establish a one-to-one mapping relationship between the examination room or secure room number and the video clips; Return to the judgment result of selecting a video segment, until all judgment results have been selected; Generate statistical event data to be fed back.

10. A system for analyzing and monitoring the behavior of examination rooms and secure rooms using intelligent agents, characterized in that: include: A server is configured to execute the examination room and confidential room behavior analysis and inspection method using intelligent agents as described in any one of claims 1 to 9; The camera device is connected to the server for video capture.

Images