Wind tunnel safety management and control system and safety management and control method based on image recognition

Abstract

The invention discloses a wind tunnel safety management and control system based on image recognition, and the system comprises a front-end intelligent safety management and control device and a rear-end intelligent safety management platform, and the front-end intelligent safety management and control device and the rear-end intelligent safety management platform are in wireless connection through a network device. The front-end intelligent safety management and control devices are arranged at the windward hole and the downwind hole respectively, and the front-end intelligent safety management and control device located at the windward hole and the front-end intelligent safety management and control device located at the downwind hole are in wireless connection through the network device. The invention further discloses a safety management and control method applied to the wind tunnel safety management and control system based on image recognition. Identification of people and articles entering and exiting from the wind tunnel of the generator, detection of metal articles carried by people, alarm management and article storage are realized; by means of article modeling, incoming and outgoing articles can be automatically counted and recorded, metal articles carried by people can be recognized, registration efficiency is improved, and the purpose of improving the safety management level of hydropower station generator wind tunnel maintenance operation is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of safety control of hydropower stations, and in particular to a wind tunnel safety control system and a safety control method based on image recognition. Background Art

[0002] During the maintenance of the hydropower station unit, a large number of workers need to enter and exit the generator wind tunnel, and they need to carry a large number of tools and maintenance materials. In order to prevent the occurrence of unsafe incidents such as generator short circuit, unit sweeping, and unit tripping due to the loss of metal objects during the generator maintenance work, the maintenance site needs to strictly register and close the personnel and their belongings entering and leaving the generator wind tunnel. At present, the entry and exit management of personnel, tools and materials in the generator wind tunnel adopts the method of manual registration and closing by external personnel on duty, which leads to the need for a large number of external personnel to perform on-duty inspection of the generator wind tunnel during the unit maintenance to ensure the safety of the maintenance. The registration and closing of personnel and items entering and leaving the generator wind tunnel is inefficient and consumes a lot of manpower costs. In order to improve the intelligent level of generator wind tunnel safety management, it is necessary to develop a wind tunnel safety management system and safety management method based on image recognition and apply it to the maintenance site of the hydropower station. Summary of the invention

[0003] The purpose of the present invention is to provide a wind tunnel safety management and control system and a safety management and control method based on image recognition, which can complete the safety monitoring of the entry and exit of personnel and objects in the generator wind tunnel during maintenance operations of a hydropower station in real time, accurately and effectively.

[0004] In order to achieve the above-mentioned purpose, the technical solution adopted by the present invention is as follows: a wind tunnel safety control system based on image recognition, including a front-end intelligent safety control device and a back-end intelligent safety management platform, the front-end intelligent safety control device and the back-end intelligent safety management platform are wirelessly connected through a network device; the front-end intelligent safety control device includes a safety control device, an object recognition device and an intelligent locker, the safety control device is integrated with a face recognition system and an object recognition system, the object recognition device is integrated with an object recognition system, and the intelligent locker is integrated with a locker management system; the front-end intelligent safety control device is respectively placed at the upper wind tunnel entrance and the lower wind tunnel entrance, and the front-end intelligent safety control device located at the upper wind tunnel entrance and the front-end intelligent safety control device located at the lower wind tunnel entrance are respectively wirelessly connected through network devices to share data.

[0005] Furthermore, the security management and control device is integrated with a face information database, an object information database, and a face model database, the object recognition device is integrated with an object information database and an object model database, and the smart locker is integrated with monitoring data.

[0006] Furthermore, the item model database is integrated with an item recognition algorithm, which includes an item segmentation and positioning algorithm, an item classification and recognition algorithm, and an image enhancement algorithm. The item recognition system recognizes items by calling the item recognition algorithm integrated in the item model database to form item information data.

[0007] Furthermore, the face model database is integrated with a face recognition algorithm, and the face recognition algorithm includes a recognition algorithm based on face feature points, a recognition algorithm based on the entire face image, a recognition algorithm based on a template, and an algorithm using a neural network for recognition. The face recognition system intelligently recognizes the face by calling the face recognition algorithm integrated in the face model database to form face information data.

[0008] Furthermore, the security control device is provided with a camera, an audio-visual alarm indicator, and a metal detector; and is equipped with an integrated control console, which integrates a face recognition system. The face recognition system is provided with a personnel management module and an alarm control module. The personnel management module can flexibly set entry and exit control personnel, time period control, and blacklist loss and addition, and perform blacklist and whitelist management in different scenarios; the alarm control module is integrated with an alarm database, which can edit alarm information, adjust alarm sound and category; the integrated control console is wirelessly connected to the background intelligent security management platform.

[0009] Furthermore, the face information database includes a face capture library, a face registration library and a blacklist library, among which: the face capture library: contains historical capture scene pictures, face thumbnails and structured face feature data, capture location, capture time information, mainly used for picture retrieval and comparison, query the target person's portrait location, time, trajectory tracking; face registration library: mainly used to import portrait pictures, structured face feature data and ID card information, after importing, it is mainly used for picture retrieval and comparison and identity information query to determine the identity of the person; blacklist library: contains unregistered and unbooked face pictures, structured face feature data and person identity information.

[0010] Furthermore, an alarm threshold is set in the alarm control module.

[0011] Furthermore, a maintenance time threshold is set on the intelligent safety management platform.

[0012] A safety control method is applied to the wind tunnel safety control system based on image recognition, comprising the following steps: S1 Personnel registration: Use security control devices to perform face recognition, record the entry information of personnel entering the venue, and form the identification information of personnel entering the venue; S2 Scanning Tools: Use the object recognition device to scan the tools carried by the entrants, identify the objects on the scanning tools, match the scanned images with the object information database in the object recognition device, and store them in the database to form the information of the objects entering the venue; S3 data synchronization: upload the on-site personnel information and on-site item information in steps S1 and S2 to the back-end intelligent security management platform in real time; S4 Exit monitoring: Collect the exit personnel and exit article information, and compare it with the entry personnel and entry article information formed in steps S1 and S2. Only when the exit personnel and exit article information are consistent with the entry personnel and entry article information can the personnel and articles leave the site, and the tools carried by the maintenance personnel shall be written off.

[0013] Furthermore, in step S1, objects that may be hidden by the person are scanned and identified, the object information is stored in a corresponding object information database, and it is identified whether the person has hidden any metal objects.

[0014] Furthermore, step S2 also includes metal detection: taking photos of the metal tools involved to confirm and generate warning information; then the on-site personnel information, on-site item information, and warning information in steps S1 and S2 are uploaded to the back-end intelligent security management platform in real time and synchronously.

[0015] Compared with the prior art, the advantages of the present invention are: 1. The present invention realizes the identification of people and objects entering and leaving the wind tunnel of the generator, the detection of metal objects carried by people, alarm management and object storage; the automatic identification and recording of people and objects and metal detection are realized when entering and leaving the wind tunnel of the generator. Through face recognition technology, the people entering and leaving can be automatically counted and recorded. Through object modeling, the objects entering and leaving can be automatically counted and recorded. It can identify metal objects carried by people and issue an alarm prompt. The staff can enter the wind tunnel only after confirmation, thereby improving the registration efficiency, so as to achieve the purpose of improving the safety management level of wind tunnel maintenance operations of hydropower station generators.

[0016] 2. The entire system realizes communication between the front-end intelligent security control equipment and the back-end intelligent security management platform through the local area network. When the network link is disconnected, it will not affect the normal personnel and item entry and exit registration and security management functions of the front-end equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 A schematic diagram of the structure of the present invention; Figure 2 An entry flow chart of the safety control device in the present invention; Figure 3 An object recognition flow chart of the object recognition device of the present invention. DETAILED DESCRIPTION

[0018] In order to make the purpose, technical scheme and advantages of the embodiments of the present invention clearer, the technical scheme 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 the embodiments. For those skilled in the art, the present invention may have various changes and variations. 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. Example 1

[0019] See also Figure 1 , a wind tunnel safety control system based on image recognition includes a front-end intelligent safety control device and a back-end intelligent safety management platform, the front-end intelligent safety control device and the back-end intelligent safety management platform are wirelessly connected through network equipment; the front-end intelligent safety control device includes a safety control device, an object recognition device and an intelligent locker, the safety control device is integrated with a face recognition system and an object recognition system, the object recognition device is integrated with an object recognition system, and the intelligent locker is integrated with a locker management system; the front-end intelligent safety control device is respectively placed at the upper wind tunnel entrance and the lower wind tunnel entrance, and the front-end intelligent safety control device located at the upper wind tunnel entrance and the front-end intelligent safety control device located at the lower wind tunnel entrance are respectively wirelessly connected to the back-end intelligent safety management platform through network equipment to share data. In this application, the network device can be a server, a switch, a wireless repeater, a router, etc.

[0020] In the present application, an integrated control console is provided in the security control device, and the integrated control console is wirelessly connected to the background intelligent security management platform. The integrated control console integrates a facial information database, an item information database, and a model database; a control center is provided in the item identification device, and the control center integrates an item information database and a model database; an industrial computer is provided in the intelligent locker, and the locker management system is integrated on the industrial computer. The locker management system includes an item management system, a locker control system, and is integrated with monitoring data. When personnel and tools enter the upwind tunnel entrance and exit the downwind tunnel entrance, the front-end intelligent security control equipment located at the upwind tunnel entrance and the front-end intelligent security control equipment located at the downwind tunnel entrance respectively scan and identify the personnel and tools, and store the extracted facial information and item information in the corresponding facial information database and item information database. In some feasible implementations, the facial information database includes a facial capture library, a facial registration library and a blacklist library, wherein: the facial capture library contains historical on-site capture pictures, facial thumbnails and structured facial feature data, capture locations, capture times and other information, and is mainly used for image retrieval and comparison, and to query the target person's portrait location, time, and trajectory tracking; the facial registration library is mainly used to import portrait pictures, structured facial feature data and ID card information, and after import it is mainly used for image retrieval and comparison and identity information query to determine the identity of the person; the blacklist library contains unregistered and unbooked facial pictures, structured facial feature data and person identity information.

[0021] The security management and control device is equipped with cameras, sound and light alarm indicators, and metal detectors; the integrated control console is integrated with a face recognition system, and the face recognition system is equipped with a personnel management module and an alarm control module. The personnel management module can flexibly set up entry and exit control personnel, time period control, and blacklist loss and addition, and blacklist and whitelist management for different scenarios; the alarm control module can integrate the alarm information with an alarm database, and can edit the alarm information, and adjust the alarm sound and category.

[0022] See also Figure 2First, a whitelist is set in the personnel management module of the face recognition system. When the maintenance personnel enter the site, the security control device scans and identifies the maintenance personnel themselves, and scans and identifies the items that the maintenance personnel may hide. When scanning and identifying the maintenance personnel themselves, the face recognition system uses distributed cluster technology, deep learning technology based on neural networks, and massive data storage big data computing technology. The face recognition system calls the recognition algorithm based on facial feature points, the recognition algorithm based on the entire face image, the recognition algorithm based on templates, and the algorithm for recognition using neural networks integrated in the face model database. Through real-time video monitoring images, video images are transmitted in the form of video streams or picture streams, and face images of the on-site environment are provided. The face is intelligently identified to determine whether it belongs to the information in the whitelist. If it belongs to the information in the whitelist, the match is successful and the face information is formed. The face information is stored in the corresponding face information database and transmitted to the back-end intelligent security management platform to record the registration information. If the information is not in the whitelist, the face information will be entered into the blacklist information. There are two ways to solve the blacklist problem. One is that the security inspector can enter the face information into the whitelist, and then pass the identification of the security control device again; the other is to perform manual registration so that the face information is no longer identified by the security control device. If the face information cannot pass the identification of the security control device and cannot pass the approval of the security inspector, it will not be allowed to enter. In the face recognition process, the face recognition system adopts a dynamically scalable distributed structure. The bottom layer is responsible for the library management of the portrait library, and handles the bottom-level applications such as portrait library creation, feature value extraction, and portrait indexing. Provide portrait 1:1 comparison and 1:N comparison, n:N comparison, dynamic comparison, etc., and realize the integration of person and certificate, static retrieval, one person with multiple certificates, whitewashing personnel identity confirmation and other big data application processing through the API interface, and quickly realize the integration of portraits on the back-end intelligent security management platform. During the process of scanning and identifying items that may be hidden by the maintenance personnel themselves, hidden items are detected and the item information is stored in the corresponding item information database; and a metal detector is used to detect whether the maintenance personnel themselves have hidden metal objects. If metal objects are found, the sound and light alarm indicator will sound an sound and light alarm. When the metal objects are taken out, the maintenance personnel will pass through the safety control device again. They can enter the wind tunnel entrance only when the metal objects are detected.

[0023] See also Figure 3After the information of the on-site maintenance personnel is approved, the maintenance personnel will put the tool on the operating table of the item identification device, and the item identification device will scan the tool. When the number of items is more than one and the item identification device can identify all items, the name and quantity of the matching items will be formed. After the security inspector confirms the item information again, the item information data will be saved and transmitted to the back-end intelligent security management platform to record the registration information. When the number of items is more than one and the item identification device cannot identify all items, the name, quantity and other information of the unidentified items will be manually registered by the security inspector. Then, after the security inspector confirms the name, quantity and other information of the matched identified items and the name, quantity and other information of the manually registered unidentified items, the item information will be saved and transmitted to the back-end intelligent security management platform to record the registration information. In the process of item identification, the item identification device retrieves the data model of the model database to model the tool on site. After the modeling is successful, the item information formed can be shared in the item identification device located at the upper wind tunnel entrance and the item identification device located at the lower wind tunnel entrance. When the size of the tool exceeds the size of the identification platform of the object identification device or cannot be identified at all, the information can be manually entered. The manually entered information may include the personnel ID photo, crew, personnel name, personnel alias, work number, department, mobile phone number, remarks and other information.

[0024] The front-end intelligent security control equipment at the upper wind tunnel entrance and the front-end intelligent security control equipment at the lower wind tunnel entrance compare the facial information and item information of the people entering and leaving the venue, and transmit the comparison results to the back-end intelligent security management platform; synchronously, the back-end intelligent security management platform will also compare the facial information and item information of the front-end intelligent security control equipment at the upper wind tunnel entrance and the front-end intelligent security control equipment at the lower wind tunnel entrance; at the same time, voice prompts and text display interfaces are output on the front-end intelligent security control equipment. It should be noted that when the front-end intelligent security control equipment is disconnected from the management platform network, it can also work independently on a single machine. When the network is disconnected, the single machine can be used to model the maintenance tool, capture pictures, and obtain the learning mode of the feature vector by calling the data model of the model database to store the face recognition and item recognition; or collect face information and item information under the administrator account, or import face information and item information data through USB, verify the identity and face data on the front-end intelligent security control equipment, and then output voice prompts and text display interfaces, and perform multi-type and multi-level authority management on personnel to deal with the entry and exit control in the control area.

[0025] When maintenance personnel enter or leave the site, they are checked by the facial recognition system and item recognition system of the safety control devices placed at the upwind tunnel entrance and the downwind tunnel entrance respectively, as well as the item recognition system of the item recognition device. Only when the facial information data and item information data are consistent, the tools carried by the maintenance personnel can be cancelled; if the tools carried by the maintenance personnel are not cancelled, the facial information data records and item information data records will be permanently recorded until the maintenance personnel's entry and exit information are consistent, forming a closed loop of cancellation.

[0026] In this application, in order to realize the ability to quickly and accurately identify the identity of personnel and item information, the personnel information and item information are originally set in the face information database and the item information database. When entering and leaving the site, the information of the face information database and the item information database is read in real time. Only verification is required. All information can be queried, and the entry and exit of personnel and items are controlled to ensure that no irrelevant personnel or items enter the generator wind tunnel. At the same time, with the permission setting, personnel can enter the corresponding places according to work needs. If illegal entry occurs, the face recognition system captures the image and basic information of the face information database, and issues a warning through voice prompts and sound and light alarms. It should be noted that the personnel information deletion settings can be performed on both the back-end intelligent security management platform and the front-end intelligent security control equipment, and support functions such as uploading photos and attachments to provide personnel information, customized functions of screening conditions, real-time alarm reminders, and synchronous management of the original setting information.

[0027] In this application, the security control device adopts an integrated security door; the security door is equipped with a camera, an audible and visual alarm indicator, and a metal detector. The integrated security door takes face recognition technology as the core. Through the camera, the facial features are identified, and the face recognition system is associated to scan and identify the personnel entering or leaving the venue, and a blacklist is established for unidentified personnel. Specifically, a 360-degree rotating camera is installed on the security door, and an integrated control console is electrically connected to the camera. The embedded Linux operating system is adopted. The integrated control console integrates the face recognition system, adopts a deep learning algorithm, and supports single or multiple person recognition. At the same time, the integrated control console has networking function, wired network, WiFi; communication methods and network protocols such as EZVIZ protocol / ISAPI / ISUP5.0, wireless connection with the background intelligent security management platform, and further wireless connection with the power station gate control system; in addition, a single machine can be connected to the Internet and parameters can be configured through the web terminal; the platform can also be used to aggregate passenger flow data and alarm data, and the face information database can be transmitted to the background intelligent security management platform to realize remote data information management.

[0028] When the wind tunnel maintenance personnel enter the integrated security door, metal detection can be performed on the wind tunnel maintenance personnel. When the metal detector identifies metal objects on the neck, waist, wrist, pocket and sole of the person, multiple metals at different positions of the human body will be alarmed at the same time when passing through the security door, which can indicate the location of multiple metals and effectively verify prohibited items. At the same time, the background intelligent security management platform or the hydropower station gate control system will also alarm at the same time. In this application, an alarm threshold is set in the alarm control module, and the alarm threshold can be adjusted according to the size of the metal to be detected, supporting 255 levels. The alarm is cancelled until the person entering confirms and enters the attribute of the item as a non-prohibited item, and the wind tunnel maintenance personnel can enter. A maintenance time threshold is set on the intelligent security management platform. When the maintenance personnel do not leave for a long time, when their stay time in the wind tunnel is greater than the preset maintenance time threshold, an alarm is issued on the background intelligent security management platform; or an alarm is issued on the background intelligent security management platform and the hydropower station gate control system at the same time. When the sound and light alarm is activated, the alarm light continues to flash until the alarm signal disappears or human intervention occurs, and the volume of the buzzer is adjustable in the alarm control module. In this application, the sound and light alarm function has an independent reset switch, and also has the function of completely turning off the sound alarm, which can be adjusted or preset in the alarm control module. The security control device of this application achieves a facial recognition accuracy of up to 99%, with a rate of no more than 3 seconds; the object image recognition rate is greater than 85%, and the object recognition rate is no more than 5 seconds. It supports batch registration and batch verification, supports networking of multiple devices, and multiple entrances and exits registration and verification functions. It greatly improves the identification and alarm work of personnel entering and leaving the venue, and improves the efficiency of registration and verification when personnel enter and leave the venue.

[0029] In the present application, the object recognition device is installed with X-ray imaging software, and includes a control center, a camera, a light source device, and an audible and visual alarm. An object information database and an object model database are set on the control center. Specifically, the object model database is integrated with an object recognition algorithm, and the object recognition algorithm includes an object segmentation and positioning algorithm, an object classification and recognition algorithm, and an image enhancement algorithm. After scanning the object placed in the object recognition device, the object recognition system identifies the object by calling the object recognition algorithm integrated in the object model database to form object information data, and stores the object information data in the corresponding object information database, and shares the object information in the object recognition device located at the upper wind tunnel entrance and the object recognition device located at the lower wind tunnel entrance. In the present application, the spatial resolution of the object camera reaches 1.0mm. The light source device adopts an X-ray generator, which uses a 160kV voltage and a 1.2mA adjustable current environment. When prohibited tools are detected passing through the object identification device, the background intelligent security management platform and the hydropower station gate control system will simultaneously sound an alarm. The alarm will not be cancelled until the security inspector confirms and enters the attributes of the item as a non-prohibited item. Only then can the wind tunnel maintenance personnel enter with the tools.

[0030] In the present application, the smart locker includes a locker body, which is composed of multiple storage boxes, and a 2 million infrared pixel camera is installed on the locker body; the locker body is made of 0.8 mm thick cold-rolled steel plate, and the locker is divided into multiple storage compartments, which can be designed in various sizes, and multiple lockers are electrically connected together. An industrial computer is set on the smart locker, and the locker management system is integrated on the industrial computer. The locker management system includes an item management system, a locker control system, and is integrated with a monitoring database. The B / S architecture is adopted to realize the remote control of the use of the smart locker in the local area network, including remote unpacking, querying usage records, report export, the use of the box door, the online status of the smart locker, etc.; and the use information of the smart locker, the liveness detection information, etc. are stored in the monitoring database; at the same time, the face recognition system in the security control device is also associated to perform face recognition and liveness detection, and can recognize live faces, effectively avoiding the opening of the locker with photos and pictures. Example 2

[0031] The safety control method applied to the wind tunnel safety control system based on image recognition comprises the following steps: S1 Personnel registration: Use security control devices to perform face recognition, record the entry information of the entrants, and form the entrant identification information; the entrant identification information mainly includes the facial features of the entrants; S2 Scanning tools: Use the item recognition device to scan the tools carried by the entrants, identify the items on the scanning tools, match the scanned images with the item information database in the item recognition device, and return them to the database to form the entry item information; the entry item information may include the item name, alias, quantity, status, state, entry time, person picture, overall picture of the item, individual picture of the item, etc. S3 data synchronization: upload the on-site personnel information and on-site item information in steps S1 and S2 to the back-end intelligent security management platform in real time; S4 Exit monitoring: Collect the exit personnel information and exit item information of the exit personnel, and compare them with the entry personnel information and entry item information formed in steps S1 and S2. Only when the exit personnel information and exit item information are consistent with the entry personnel information and entry item information can the personnel and items leave the site, and the tools carried by the maintenance personnel shall be written off.

[0032] Specifically, after the personnel entering the site register through the facial recognition system and the object recognition system of the security control device, they will place the maintenance tools they carry on the object recognition device and start the object recognition device for object recognition. If there are more than one maintenance tool, they need to be placed one by one. It takes 3 seconds to identify each maintenance tool. The maintenance tools will be identified one by one, and the identified pictures, quantity, and total number of maintenance tools will be displayed separately, and the on-duty personnel will check and cancel them.

[0033] In some feasible implementations, in step S1, items that may be hidden by the person are scanned and identified, the item information is stored in the corresponding item information database, and it is identified whether the person is hiding metal items to generate warning information. Then, in step S3, the information on the person entering the site, the hidden item information or the warning information in step S1 is uploaded to the back-end intelligent security management platform in real time and synchronously.

[0034] In some feasible implementations, step S2 also includes metal detection: taking photos of the metal tools involved to generate warning information; and then in step S3, the information on incoming items or the warning information in step S2 is uploaded to the back-end intelligent security management platform in real time.

Claims

1. A wind tunnel safety control system based on image recognition, characterized in that: It includes front-end intelligent security control equipment and back-end intelligent security management platform, which are wirelessly connected to each other through network equipment; the front-end intelligent security control equipment includes a security control device, an item recognition device and an intelligent locker, the security control device is integrated with a face recognition system and an item recognition system, the item recognition device is integrated with an item recognition system, and the intelligent locker is integrated with a locker management system; the front-end intelligent security control equipment is respectively placed at the upper wind tunnel entrance and the lower wind tunnel entrance, and the front-end intelligent security control equipment located at the upper wind tunnel entrance and the front-end intelligent security control equipment located at the lower wind tunnel entrance are wirelessly connected to each other through network equipment to share data.

2. The wind tunnel safety management and control system based on image recognition according to claim 1, characterized in that: The security control device is integrated with a face information database, an item information database, and a face model database; the item recognition device is integrated with an item information database and an item model database; and the smart locker is integrated with monitoring data.

3. The wind tunnel safety management and control system based on image recognition according to claim 2 is characterized in that: The item model database is integrated with an item recognition algorithm, which includes an item segmentation and positioning algorithm, an item classification and recognition algorithm, and an image enhancement algorithm. The item recognition system recognizes items by calling the item recognition algorithm integrated in the item model database to form item information data.

4. The wind tunnel safety management and control system based on image recognition according to claim 2 is characterized in that: The face model database is integrated with a face recognition algorithm, which includes a recognition algorithm based on face feature points, a recognition algorithm based on the entire face image, a recognition algorithm based on a template, and an algorithm that uses a neural network for recognition. The face recognition system intelligently recognizes the face by calling the face recognition algorithm integrated in the face model database to form face information data.

5. The wind tunnel safety management and control system based on image recognition according to claim 1, characterized in that: The safety control device is provided with a camera, an audible and visual alarm indicator, and a metal detector; and is equipped with an integrated control console, which integrates a face recognition system, a face recognition system It is equipped with a personnel management module and an alarm control module. The personnel management module can flexibly set up entry and exit control personnel, time period control, and blacklist loss and addition, and perform blacklist and whitelist management in different scenarios; the alarm control module is integrated with an alarm database, which can edit alarm information, adjust alarm sound and category; the integrated control console is wirelessly connected to the background intelligent security management platform.

6. The wind tunnel safety management and control system based on image recognition according to claim 2 is characterized in that: The face information database includes the face capture database, face registration database and blacklist database, among which: Face capture database: contains historical capture scene pictures, face thumbnails and structured face feature data, capture location, and capture time information. It is mainly used for image retrieval and comparison, querying the location, time, and trajectory tracking of the target person's portrait; Face registration library: mainly used to import portrait images, structured facial feature data and ID card information. After importing, it is mainly used for image retrieval and comparison and identity information query to determine the identity of the person; Blacklist library: contains facial images of unregistered and unbooked people, structured facial feature data, and personal identity information.

7. The wind tunnel safety management and control system based on image recognition according to claim 5 is characterized in that: An alarm threshold is set in the alarm control module.

8. The wind tunnel safety management and control system based on image recognition according to claim 1, characterized in that: A maintenance time threshold is set on the intelligent safety management platform.

9. A safety control method, applied to the wind tunnel safety control system based on image recognition according to any one of claims 1 to 8, characterized in that: The following steps are involved: S1 Personnel registration: Use security control devices to perform face recognition, record the entry information of personnel entering the venue, and form the identification information of personnel entering the venue; S2 Scanning Tools: Use the object recognition device to scan the tools carried by the entrants, identify the objects on the scanning tools, match the scanned images with the object information database in the object recognition device, and store them in the database to form the information of the objects entering the venue; S3 data synchronization: upload the on-site personnel information and on-site item information in steps S1 and S2 to the back-end intelligent security management platform in real time; S4 Exit monitoring: Collect the exit personnel and exit article information, and compare it with the entry personnel and entry article information formed in steps S1 and S2. Only when the exit personnel and exit article information are consistent with the entry personnel and entry article information can the personnel and articles leave the site, and the tools carried by the maintenance personnel shall be written off.

10. The security management method according to claim 9, characterized in that : In step S1, the objects that may be hidden by the person are scanned and identified, the object information is stored in the corresponding object information database, and it is identified whether the person has hidden metal objects.

11. The security management and control method according to claim 9, characterized in that: Step S2 also includes metal detection: taking photos of the metal tools involved to form warning information; then uploading the on-site personnel information, on-site item information, and warning information in steps S1 and S2 to the back-end intelligent security management platform in real time.

Images