Workshop safety management system and method

Abstract

The workshop safety management system comprises an AI terminal, a network camera, a video switch, a mobile phone WeChat terminal and a cloud server, the AI terminal is connected with the video switch, the network camera is connected with the video switch, and the AI terminal obtains a plurality of network camera video streams through the network switch and sends the video streams to the mobile phone WeChat terminal. The safety management system is connected to the cloud server through an MQTT protocol or a vpn special line, the cloud server can perform data interaction with the mobile terminal WeChat, the AI terminal carries a backpack recognition algorithm, a work clothes recognition algorithm and a face analysis algorithm, and the AI terminal is provided with a web management page. The web management page can be configured to access network camera monitoring point location information for video analysis and associated access control authentication AI terminal ID information, and the access control system comprises an access controller and an authentication terminal; according to the scheme, the image frame rate capable of being processed by each video monitoring point is dynamically adjusted, so that the system performance is exerted to the maximum extent, and the optimal recognition and analysis effect is achieved.

Description

Technical Field

[0001] The present invention belongs to the technical field of workshop management, and in particular relates to a workshop safety management system and method. Background Art

[0002] Through the intelligent system, real-time monitoring and analysis of parks, workshops, warehouses and other areas can be achieved, which can effectively improve management efficiency and the safety of production operations.

[0003] Chinese invention patent CN116434455A discloses a silk-making workshop safety management method and a silk-making workshop safety management system, which includes: obtaining the photoelectric signal generated by the early warning mechanism, judging whether there is a person entering the bottom area of ​​the turned-up tobacco basket according to the preset rules, and if so, generating a first alarm and a control instruction for controlling the action of the basket turning machine; obtaining the temperature information collected by the infrared acquisition device arranged near the flue, judging whether the flue temperature exceeds the threshold according to the temperature information, and if so, generating a second alarm; obtaining the image data collected by the image acquisition device arranged in the high-risk working area, using the image data as input to judge whether there is a person who does not wear a safety helmet in the high-risk working area according to the preset image recognition model, and if so, generating a third alarm. The use of this method can expand the scope of the inspection period, improve the timeliness and effectiveness of the inspection, reduce the labor intensity of the inspectors, and achieve the goal of improving management efficiency. However, the means proposed by the aforementioned scheme are relatively single, lack of data interaction and linkage control between systems, and the resource allocation within the system is relatively fixed, which is not conducive to maximizing the performance of resources.

[0004] Therefore, the present invention provides a workshop safety management system and method for solving the problems raised by the above-mentioned background technology. Summary of the invention

[0005] In view of the problems raised by the above background technology, the purpose of the present invention is to provide a workshop safety management system and method. In combination with the application scenario that some areas in the workshop are not allowed to enter due to material storage or other management requirements, this scheme proposes a production safety management system based on video analysis and its linkage control method, and proposes a method that can dynamically adjust the image frame rate that can be processed by each video monitoring point according to the existing system resources and the recognition function to be completed, so as to maximize the system performance and achieve the best recognition and analysis effect; by identifying whether people in different areas have violated the rules, different response measures are taken, such as real-time voice broadcast reminders, or data interaction with the access control management system, and the access control management function is used to forcibly intercept violators outside the entry area to ensure the safety of property in the area.

[0006] In order to achieve the above technical objectives, the technical solution adopted by the present invention is as follows:

[0007] A workshop safety management system includes an AI terminal, a network camera, a video switch, a mobile phone WeChat terminal and a cloud server, wherein the AI ​​terminal is connected to the video switch, the network camera is connected to the video switch, the AI ​​terminal obtains multiple network camera video streams through the network switch, the safety management system accesses the cloud server through the MQTT protocol or a VPN dedicated line, the cloud server can interact with the mobile WeChat for data, the AI ​​terminal is equipped with a backpack recognition algorithm, a work clothes recognition algorithm and a face analysis algorithm, the AI ​​terminal supports a TTS function, the AI ​​terminal is connected to the network camera through the network, the AI ​​terminal is provided with a web management page, the web management page can be configured to access the network camera monitoring point information for video analysis and the associated access control authentication AI terminal ID information, and also includes an access control management system, the access control management system includes an access control controller and an authentication terminal.

[0008] It is further defined that the AI ​​terminal module has at least two independent network ports, one of which is connected to the video switch and the other is connected to the external network.

[0009] It is further defined that the network camera is provided with voice broadcast and voice intercom functions, the AI ​​terminal sends the voice reminder information to the network camera, and the network camera performs voice broadcast according to the voice reminder content.

[0010] A workshop safety management method comprises the following steps:

[0011] Step 1: Set up a security management system, access control management system and mobile location terminal. The security management system includes AI terminals, network cameras, video switches, mobile WeChat terminals and cloud servers. The access control management system includes access controllers and authentication terminals.

[0012] Step 2: When the AI ​​terminal detects a violation, it sends a corresponding voice reminder to the network camera, which then broadcasts the reminder. The web management page set by the AI ​​terminal can be used to configure the camera monitoring point information for video analysis and the associated access control authentication terminal ID information;

[0013] At the same time, when the AI ​​terminal detects a violation, the security management system sends the ID information of the access control authentication terminal at the monitoring point and whether the authentication is allowed to pass to the access control management system. The access control management system controls whether the authentication terminal is allowed to pass by judging the authentication terminal ID and its valid status;

[0014] Step 3: The AI ​​terminal estimates the number of video streams that are still allowed to be accessed based on the configured access recognition algorithm function, and can be viewed on the web or WeChat mobile app;

[0015] Step 4: Determine the number of AI terminal video streams, and divide the video monitoring area into the first area and the second area according to the strictness of the area management. Each area has an entrance and exit to the area.

[0016] Specifically, when no violation is detected at the entrances and exits of each area, the entrance and exit of the first area is the first scene where the target appears, and the image frames with the best recognition effect should be used for analysis and processing in this area. The video monitoring points in other areas can automatically adjust the number of image frames to be processed according to whether the target event is detected at the entrance and exit of the first area. When a violation is detected at the entrance and exit of the first area, the AI ​​terminal performs maximum frame extraction and recognition analysis on all video monitoring points in the first area and the video streams of the monitoring cameras at the entrance and exit of the second area. The remaining monitoring points calculate the maximum number of image frames that can be processed per second at each monitoring point video stream based on the remaining reasoning capability of the terminal and the number of cameras.

[0017] Step 5: Set a virtual wireframe for judging the entry and exit of people at the video monitoring points in each area. The virtual wireframe defines the direction information of the monitoring area. The AI ​​terminal can judge whether the illegal target person enters or leaves the monitoring area based on the target tracking algorithm and the moving direction of the illegal target person;

[0018] Step 6: When the target person violating the rules passes through the entrance and exit of the first area and is about to enter the first area, the security management system detects the target person violating the rules and issues a voice reminder not to allow the target person violating the rules to enter the second area;

[0019] When the security management system detects that an illegal target person enters the entrance and exit of the first area, the security management system generates a corresponding event record and pushes it to the administrator's mobile phone WeChat terminal in real time through the network. The administrator can view the event video n seconds before and after the event through the mobile phone WeChat terminal, and conduct real-time video viewing and remote voice intercom as needed. After the processing is completed, the event processing confirmation is completed on the mobile phone WeChat terminal. In addition, the security management system dynamically captures the characteristic information of the illegal target person through the network camera, and the characteristic information includes the facial characteristic information of the illegal target person, and creates a temporary record of the acquired characteristic information and saves it;

[0020] When the security management system detects that a target person who violates the rules has entered the first area, the security management system dynamically captures the feature information of the target person who violates the rules through a network camera, and compares and analyzes it with the facial feature information in the temporary record to track and locate the position of the person;

[0021] Step 7: When the security management system detects that a violative target person enters the entrance / exit of the second area, the security management system obtains the real-time video stream through the network camera at this entrance / exit, collects the facial features of the violative target person, and compares and analyzes them with the facial feature information in the temporary record to achieve the tracking and positioning of the person's location; each time a violative person is tracked, a corresponding event record will be generated, the camera information for real-time video viewing and remote intercom will be updated, and it will be pushed to the security administrator. The security administrator can quickly view the on-site situation through the above-pushed event information;

[0022] When the violative target person enters the second area through the entrance / exit of the second area, the access control authentication terminal bound to the video monitoring point in this area will cause the current access control to prohibit passage due to the nearby violative target person. The security management system issues a reminder through the camera that the access control is temporarily not passable due to the violative target person, and asks the violative target person to leave as soon as possible; when the security management system detects that the violative target person leaves the area set by the video monitoring point, the security management system sends information to the access control management module to restore the authentication function of the authentication terminal in this area. After receiving this information, the access control management module restores the access control authentication function at this point;

[0023] Step 8: When the security management system detects that the violative target person leaves from the first area or the second area, the security management system generates a corresponding alarm event and deletes the record. According to the location and entry / exit time information of this person's successive appearances, a corresponding violation report is generated and pushed to the mobile WeChat end of the security administrator. The security administrator can view the corresponding event records and reports on the mobile WeChat end to timely grasp the relevant dynamics.

[0024] A workshop security management method. The recognition algorithm for determining the number of video stream paths a in Step 4 adopts target tracking or target detection. When single-channel video stream target detection requires m fps images and target tracking requires n fps images, and m < n; then when this channel of video stream only requires the target detection function, the maximum number of video streams that can be accessed is 100 / m; if this channel of video stream only requires the target tracking function, the maximum number of video streams that can be accessed is 100 / n; because m < n, so (100 / m) > (100 / n), that is, the number of video stream paths a that the terminal can access is between (100 / n) and (100 / m), that is, (100 / n) < a < (100 / m).

[0025] The beneficial effects of the present invention:

[0026] 1. This solution provides a workshop safety management system and method, which is combined with existing practical problems: some areas in the workshop do not allow potential illegal target personnel to enter due to material storage or other management requirements; this solution proposes a method that can dynamically adjust the image frame rate that can be processed by each video monitoring point according to the existing safety management system resources and the recognition functions that need to be completed, so as to maximize the system performance and achieve the best recognition and analysis effect; by identifying whether people in different areas have violated the rules, different response measures can be taken, such as real-time voice broadcast reminders, or data interaction with the access control management system, and the access control management function can be used to forcibly intercept illegal personnel outside the entry area to ensure the safety of property in the area.

[0027] 2. Once the number of video streams accessed by the AI ​​terminal of the security management system is determined, the AI ​​terminal can dynamically adjust the image processing frame rate of the monitoring video in different areas according to the different functional requirements of video analysis in different areas at different times, so as to make full use of the terminal reasoning resources and achieve the best video analysis effect; similarly, when there are multiple AI terminals in the same network and perform video analysis on multiple video streams, the security management system can coordinate the resources of all terminals and dynamically allocate and reasonably utilize the system resources of multiple AI terminals according to the number of video stream accesses and video analysis functions configured for each terminal. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] The present invention can be further illustrated by the non-limiting examples given in the accompanying drawings;

[0029] Figure 1 A system module connection diagram of a workshop safety management system and method embodiment of the present invention;

[0030] Figure 2 A video detection schematic diagram of a workshop safety management system and method embodiment of the present invention. DETAILED DESCRIPTION

[0031] In order to enable those skilled in the art to better understand the present invention, the technical solution of the present invention is further described below in conjunction with the accompanying drawings and embodiments. The technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in the field without creative work are within the scope of protection of the present invention.

[0032] It should be noted that all directional indications in the embodiments of the present invention (such as up, down, left, right, front, back, etc.) are only used to explain the relative position relationship, movement status, etc. between the components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication will also change accordingly.

[0033] In addition, the descriptions of "first", "second", etc. in the present invention are only used for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In addition, the technical solutions between the various embodiments can be combined with each other, but it must be based on the ability of ordinary technicians in the field to implement. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0034] like Figure 1-Figure 2 As shown, a workshop safety management system of the present invention includes an AI terminal, a network camera, a video switch, a mobile phone WeChat terminal and a cloud server, the AI ​​terminal is connected to the video switch, the network camera is connected to the video switch, the AI ​​terminal obtains multiple network camera video streams through the network switch, the safety management system accesses the cloud server through the MQTT protocol or a VPN dedicated line, the cloud server can interact with the mobile WeChat for data, the AI ​​terminal is equipped with a backpack recognition algorithm, a work clothes recognition algorithm and a face analysis algorithm, the AI ​​terminal supports the TTS function, the AI ​​terminal is connected to the network camera through the network, the AI ​​terminal is provided with a web management page, the web management page can be configured to access the network camera monitoring point information for video analysis and the associated access control authentication AI terminal ID information, and also includes an access control management system, the access control management system includes an access control controller and an authentication terminal.

[0035] It should be noted that when a violation is detected, the AI ​​terminal sends the corresponding voice reminder information to the network camera, which completes the voice broadcast reminder, and then configures the camera monitoring point information and associated access control authentication terminal ID information for video analysis through the AI ​​terminal's web management page. When the AI ​​terminal detects a violation, the security management system sends the access control authentication terminal ID information at the monitoring point and whether authentication is allowed to pass to the access control management system. The access control management system controls whether the authentication terminal is allowed to pass by judging the authentication terminal ID and its effective status.

[0036] Preferably, the AI ​​terminal module has at least two independent network ports, one of which is connected to the video switch and the other is connected to the external network.

[0037] Preferably, the network camera is provided with voice broadcast and voice intercom functions, the AI ​​terminal sends the voice reminder information to the network camera, and the network camera performs voice broadcast according to the voice reminder content.

[0038] A workshop safety management method comprises the following steps:

[0039] Step 1: Set up a security management system, access control management system and mobile location terminal. The security management system includes AI terminals, network cameras, video switches, mobile WeChat terminals and cloud servers. The access control management system includes access controllers and authentication terminals.

[0040] Step 2: When the AI ​​terminal detects a violation, it sends a corresponding voice reminder to the network camera, which then broadcasts the reminder. The web management page set by the AI ​​terminal can be used to configure the camera monitoring point information for video analysis and the associated access control authentication terminal ID information;

[0041] At the same time, when the AI ​​terminal detects a violation, the security management system sends the ID information of the access control authentication terminal at the monitoring point and whether the authentication is allowed to pass to the access control management system. The access control management system controls whether the authentication terminal is allowed to pass by judging the authentication terminal ID and its valid status;

[0042] Step 3: The AI ​​terminal estimates the number of video streams that are still allowed to be accessed based on the configured access recognition algorithm function, and can be viewed on the web or WeChat mobile app;

[0043] like Figure 2 As shown, step 4: determine the number of AI terminal video streams, divide the video monitoring area into the first area and the second area according to the strictness of the area management, and each area has an entrance and exit to the area;

[0044] Specifically, when no violation is detected at the entrances and exits of each area, the entrance and exit of the first area is the first scene where the target appears, and the image frames with the best recognition effect should be used for analysis and processing in this area. The video monitoring points in other areas can automatically adjust the number of image frames to be processed according to whether the target event is detected at the entrance and exit of the first area. When a violation is detected at the entrance and exit of the first area, the AI ​​terminal performs maximum frame extraction and recognition analysis on all video monitoring points in the first area and the video streams of the monitoring cameras at the entrance and exit of the second area. The remaining monitoring points calculate the maximum number of image frames that can be processed per second at each monitoring point video stream based on the terminal's remaining reasoning capability and the number of cameras. It should be noted that the maximum frame extraction and recognition analysis is determined by the algorithm with the highest requirement for image frame rate.

[0045] Step 5: Set a virtual wireframe for judging the entry and exit of people at the video monitoring points in each area. The virtual wireframe defines the direction information of the monitoring area. The AI ​​terminal can judge whether the illegal target person enters or leaves the monitoring area based on the target tracking algorithm and the moving direction of the illegal target person;

[0046] Step 6: When the target person violating the rules passes through the entrance and exit of the first area and is about to enter the first area, the security management system detects the target person violating the rules and issues a voice reminder not to allow the target person violating the rules to enter the second area;

[0047] When the security management system detects that an illegal target person enters the entrance and exit of the first area, the security management system generates a corresponding event record and pushes it to the administrator's mobile phone WeChat terminal in real time through the network. The administrator can view the event video n seconds before and after the event through the mobile phone WeChat terminal, and conduct real-time video viewing and remote voice intercom as needed. After the processing is completed, the event processing confirmation is completed on the mobile phone WeChat terminal. In addition, the security management system dynamically captures the characteristic information of the illegal target person through the network camera, and the characteristic information includes the facial characteristic information of the illegal target person, and creates a temporary record of the acquired characteristic information and saves it;

[0048] When the security management system detects that a target person who violates the rules has entered the first area, the security management system dynamically captures the feature information of the target person who violates the rules through a network camera, and compares and analyzes it with the facial feature information in the temporary record to track and locate the position of the person;

[0049] Step 7: When the security management system detects that the target person who violates the rules enters the entrance and exit of the second area, the security management system obtains the real-time video stream through the network camera at the entrance and exit, collects the facial features of the target person who violates the rules, and compares and analyzes the facial feature information in the temporary record to track and locate the person; each time a violator is tracked, a corresponding event record will be generated, the camera information for real-time video viewing and remote intercom will be updated, and pushed to the security administrator, who can quickly view the on-site situation through the above-mentioned pushed event information;

[0050] When the illegal target person enters the second area through the entrance and exit of the second area, the access control authentication terminal bound to the video monitoring point in the area will cause the current access control to be prohibited due to the nearby illegal target person, and the security management system will issue a reminder through the camera that the access control is temporarily blocked due to the illegal target person, and ask the illegal target person to leave as soon as possible; when the security management system detects that the illegal target person has left the area set by the video monitoring point, the security management system sends a message to the access control management module to restore the authentication function of the authentication terminal in the area. After receiving the message, the access control management module restores the access control authentication function of the point;

[0051] Step 8: When the security management system detects that the target person in violation has left the first area or the second area, the security management system generates a corresponding alarm event and deletes the record. According to the location and entry and exit time information of the person, a corresponding violation report is generated and pushed to the security administrator's mobile WeChat terminal. The security administrator can view the corresponding event records and reports on the mobile WeChat terminal to keep abreast of relevant developments.

[0052] The AI ​​terminal can estimate the number of video streams that are still allowed to access based on the configured access recognition algorithm function, which can be viewed on the web or mobile WeChat, allowing users to understand the usage of system resources.

[0053] Theoretically, target recognition and tracking require more images to be inferred than target recognition detection. For example, when only target detection is required for a single video stream, at least one frame of image must be inferred to obtain the inference result, while target tracking requires at least two frames to locate the position change of the target. In practical applications, the number of processed image frames will be appropriately increased according to the effect of target detection and target tracking. When the number of processed frames of a single video image is increased, the number of video streams that can be processed per second will decrease. Assuming that the target detection algorithm model of the AI ​​terminal runs on a limited hardware platform and its inference time is 10ms per frame, the terminal has the inference capability to process 100 frames of images per second.

[0054] Preferably, the recognition algorithm for determining the number of video streams a in step 4 adopts object tracking or object detection. When object detection of a single video stream requires m fps images and object tracking requires n fps images, and m < n; then when this video stream only requires the object detection function, the maximum number of video streams that can be accessed is 100 / m; if this video stream only requires the object tracking function, the maximum number of video streams that can be accessed is 100 / n; because m < n, so (100 / m) > (100 / n), that is, the number of video streams a that the terminal can access is between (100 / n) and (100 / m), that is, (100 / n) < a < (100 / m).

[0055] After the number of video streams accessed by the terminal is determined, the terminal can dynamically adjust the image processing frame rate of the monitored videos in different regions according to the different functional requirements for video analysis in different regions at different times, so as to make full use of the terminal inference resources and achieve the optimal video analysis effect.

[0056] The above embodiments only exemplarily illustrate the principles and effects of the present invention, rather than limiting the present invention. Any person familiar with this technology can modify or change the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or changes completed by those with ordinary knowledge in the technical field without departing from the spirit and technical idea disclosed by the present invention should still be covered by the claims of the present invention.

Claims

1. A workshop safety management system, characterized by: It includes an AI terminal, a network camera, a video switch, a mobile phone WeChat terminal and a cloud server. The AI ​​terminal is connected to the video switch, the network camera is connected to the video switch, the AI ​​terminal obtains multiple network camera video streams through the network switch, the security management system accesses the cloud server through the MQTT protocol or VPN dedicated line, the cloud server can interact with the mobile WeChat for data, the AI ​​terminal is equipped with a backpack recognition algorithm, a work clothes recognition algorithm and a face analysis algorithm, the AI ​​terminal supports the TTS function, the AI ​​terminal is connected to the network camera through the network, the AI ​​terminal is provided with a web management page, the web management page can be configured to access the network camera monitoring point information for video analysis and the associated access control authentication AI terminal ID information, and also includes an access control management system, the access control management system includes an access control controller and an authentication terminal.

2. A workshop safety management system according to claim 1, characterized in that: The AI ​​terminal module has at least two independent network ports, one of which is connected to the video switch and the other is connected to the external network.

3. A workshop safety management system according to claim 1, characterized in that: The network camera is provided with voice broadcast and voice intercom functions. The AI ​​terminal sends the voice reminder information to the network camera, and the network camera performs voice broadcast according to the voice reminder content.

4. A workshop safety management method according to claim 1, characterized in that: The steps include: Step 1: Set up a security management system, access control management system and mobile location terminal. The security management system includes AI terminals, network cameras, video switches, mobile WeChat terminals and cloud servers. The access control management system includes access controllers and authentication terminals. Step 2: When the AI ​​terminal detects a violation, it sends a corresponding voice reminder to the network camera, which then broadcasts the reminder. The web management page set by the AI ​​terminal can be used to configure the camera monitoring point information for video analysis and the associated access control authentication terminal ID information; At the same time, when the AI ​​terminal detects a violation, the security management system sends the ID information of the access control authentication terminal at the monitoring point and whether the authentication is allowed to pass to the access control management system. The access control management system controls whether the authentication terminal is allowed to pass by judging the authentication terminal ID and its valid status; Step 3: The AI ​​terminal estimates the number of video streams that are still allowed to be accessed based on the configured access recognition algorithm function, and can be viewed on the web or WeChat mobile app; Step 4: Determine the number of AI terminal video streams, and divide the video monitoring area into the first area and the second area according to the strictness of the area management. Each area has an entrance and exit to the area. Specifically, when no violation is detected at the entrances and exits of each area, the entrance and exit of the first area is the first scene where the target appears, and the image frames with the best recognition effect should be used for analysis and processing in this area. The video monitoring points in other areas can automatically adjust the number of image frames to be processed according to whether the target event is detected at the entrance and exit of the first area. When a violation is detected at the entrance and exit of the first area, the AI ​​terminal performs maximum frame extraction and recognition analysis on all video monitoring points in the first area and the video streams of the monitoring cameras at the entrance and exit of the second area. The remaining monitoring points calculate the maximum number of image frames that can be processed per second at each monitoring point video stream based on the remaining reasoning capability of the terminal and the number of cameras. Step 5: Set a virtual wireframe for judging the entry and exit of people at the video monitoring points in each area. The virtual wireframe defines the direction information of the monitoring area. The AI ​​terminal can judge whether the illegal target person enters or leaves the monitoring area based on the target tracking algorithm and the moving direction of the illegal target person; Step 6: When the target person violating the rules passes through the entrance and exit of the first area and is about to enter the first area, the security management system detects the target person violating the rules and issues a voice reminder not to allow the target person violating the rules to enter the second area; When the security management system detects that an illegal target person enters the entrance and exit of the first area, the security management system generates a corresponding event record and pushes it to the administrator's mobile phone WeChat terminal in real time through the network. The administrator can view the event video n seconds before and after the event through the mobile phone WeChat terminal, and conduct real-time video viewing and remote voice intercom as needed. After the processing is completed, the event processing confirmation is completed on the mobile phone WeChat terminal. In addition, the security management system dynamically captures the characteristic information of the illegal target person through the network camera, and the characteristic information includes the facial characteristic information of the illegal target person, and creates a temporary record of the acquired characteristic information and saves it; When the security management system detects that a target person who violates the rules has entered the first area, the security management system dynamically captures the feature information of the target person who violates the rules through a network camera, and compares and analyzes it with the facial feature information in the temporary record to track and locate the position of the person; Step 7: When the security management system detects that the target person who violates the rules enters the entrance and exit of the second area, the security management system obtains the real-time video stream through the network camera at the entrance and exit, collects the facial features of the target person who violates the rules, and compares and analyzes the facial feature information in the temporary record to track and locate the person; each time a violator is tracked, a corresponding event record will be generated, the camera information for real-time video viewing and remote intercom will be updated, and pushed to the security administrator, who can quickly view the on-site situation through the above-mentioned pushed event information; When a target person violating regulations enters the second area through the entrance / exit of the second area, the access control authentication terminal bound to the video monitoring point in this area will cause the current access control to prohibit passage due to the nearby target person violating regulations. The security management system sends a reminder through the camera that the access control is temporarily unavailable due to the presence of a target person violating regulations, and asks the target person violating regulations to leave as soon as possible; when the security management system detects that the target person violating regulations leaves the area set by the video monitoring point, the security management system sends information to the access control management module to restore the authentication function of the authentication terminal in this area. After receiving this information, the access control management module restores the access control authentication function of this point; Step 8: When the security management system detects that the target person violating regulations leaves the first area or the second area, the security management system generates a corresponding alarm event and deletes the record. According to the location where the person appeared successively and the entry / exit time information, a corresponding violation report is generated and pushed to the mobile WeChat terminal of the security administrator. The security administrator can view the corresponding event records and reports on the mobile WeChat terminal to timely grasp the relevant dynamics.

5. A workshop safety management method according to claim 4, characterized in that: The recognition algorithm for determining the number of video streams a in step 4 uses target tracking or target detection. When single-channel video stream target detection requires m fps images and target tracking requires n fps images, and m < n; then when only the target detection function is required for this video stream, the maximum number of video streams that can be accessed is 100 / m; if only the target tracking function is required for this video stream, the maximum number of video streams that can be accessed is 100 / n; because m < n, so (100 / m) > (100 / n), that is, the number of video streams a that the terminal can access is between (100 / n) and (100 / m), that is, (100 / n) < a < (100 / m).

Images