Mine disaster emergency escape practical training system

Abstract

The application provides a mine disaster emergency escape practical training system, and relates to the technical field of mine practical training. The system comprises a roadway simulation module, a preparation area module, a self-rescuer blind wearing practical training module, a self-rescuer AI visual identification examination system, an examination confirmation module and a self-rescuer training management system. The roadway simulation module is used for making a simulated roadway environment according to an actual roadway structure and simulating a roadway disaster environment. The preparation area module is located in the simulated roadway and is used for identifying and registering reference personnel information. The self-rescuer blind wearing practical training module is located in the simulated roadway and is used for reference personnel to wear a self-rescuer for practical training. The examination confirmation module is located in the simulated roadway and is used for examining the result of the reference personnel wearing the self-rescuer, returning the result and time to the self-rescuer training management system, confirming the examination time and scoring according to the completion time. The system can realize self-rescuer wearing training examination in a simulated disaster environment.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of mine training, in particular to a mine disaster emergency escape training system. BACKGROUND

[0002] In various types of mine disasters, the self-rescuer is the "last life barrier" for miners to realize personal escape and self-rescue in emergency situations (such as fire, explosion, and toxic and harmful gas outflow, etc.), and its correct and skilled use is crucial. However, historical experience shows that improper use or unskilled operation of the self-rescuer is one of the important reasons for the expansion of accidents and casualties.

[0003] In order to effectively strengthen the safety management of mine self-rescuers and improve the emergency self-rescue ability of employees, a roadway and system for simulating disaster environments are needed for self-rescuer wearing training and examination in simulated disaster environments (collapse, fire, smoke, etc.). SUMMARY

[0004] The technical problem to be solved by the present application is to provide a mine disaster emergency escape training system for mine disaster emergency escape training in view of the shortcomings of the prior art.

[0005] To solve the above technical problems, the technical solution adopted by the present application is: a mine disaster emergency escape training system, comprising a roadway simulation module, a preparation area module, a self-rescuer blind wearing training module, a self-rescuer AI visual identification examination system, an examination confirmation module and a self-rescuer training management system.

[0006] The roadway simulation module is used to simulate the roadway disaster environment according to the actual roadway structure to make a simulated roadway environment.

[0007] The preparation area module is located in the simulated roadway and is used for identification and registration of personnel information.

[0008] The self-rescuer blind wearing training module is located in the simulated roadway and is used for self-rescuer wearing training of reference personnel.

[0009] The examination confirmation module is located in the simulated roadway and is used to examine the results of the reference personnel wearing the self-rescuer, return the results and time to the self-rescuer training management system, confirm the examination time and score according to the completion time.

[0010] Further, the preparation area module includes an identity recognition system and a health detection system for identity recognition and health detection of reference personnel information.

[0011] Further, the identity recognition system comprises an iris recognition verification device, a face recognition device, a card swiping registration device and an alcohol detection device. After the reference personnel enter the training site preparation area, the iris recognition verification device and the face recognition device are used to verify the identity, then the card swiping registration device is used to register the number of the digital number card with RFID, and the alcohol detection device is used for alcohol detection.

[0012] Further, the health detection system adopts a smart watch to monitor the physiological index data of the reference personnel in the simulated roadway during the practical operation examination process in real time, and analyzes the physiological reaction intensity of the reference personnel in the simulated roadway disaster environment, so as to assist in evaluating the psychological stress resistance and emergency adaptability of the reference personnel. All physiological index data, student identity information, operation video and examination results are automatically bound and stored to form a complete training file.

[0013] Further, the self-rescuer blind wearing training module is provided with a self-rescuer with a detachable mouthpiece.

[0014] Further, the self-rescuer blind wearing training module is provided with a self-rescuer with a detachable mouthpiece.

[0015] Further, the self-rescuer blind wearing training module is provided with a self-rescuer with a detachable mouthpiece.

[0016] Further, the self-rescuer blind wearing training module is provided with a self-rescuer with a detachable mouthpiece.

[0017] The self-rescuer AI visual recognition examination system adopts a multi-task cooperative detection architecture, integrates three functions of human body target detection, number plate recognition and self-rescuer wearing state detection, and realizes intelligent binding of personnel identity and self-rescuer wearing state through a unique miner number plate.

[0018] The self-rescuer AI visual recognition assessment system adopts an improved yolov5 model for human target detection, number plate recognition and self-rescuer wearing state detection.

[0019] In the feature extraction stage, a multi-scale smoke robust feature extraction module is designed; the multi-scale smoke robust feature extraction module is constructed based on depth separable convolution with four-layer pyramid structure with step two for downsampling, each layer of features generates a channel weight vector through an SE attention module, and the smoke sensitive channel is suppressed through an experiment calibrated threshold function, and the key features are retained.

[0020] The multi-scale smoke robust feature extraction module is embedded in the backbone network of the yolov5 model to ensure the stability of the detection.

[0021] Further, the assessment confirmation module is provided with a display screen connected with the self-rescuer training management system, and the results of scoring and the information shot by the AI camera are displayed on the display screen.

[0022] Further, the simulated roadway further comprises a compressed air self-rescue training area and an extreme condition simulation area, wherein the compressed air self-rescue training area is used for developing compressed air self-rescue device training, and a one-key pop-up compressed air self-rescue device is arranged to realize rapid wearing and automatic air supply.

[0023] The extreme condition simulation area continuously injects carbon dioxide gas into the closed area of the simulated roadway through a carbon dioxide compressed steel cylinder, and sets up a single exhaust system to maintain the positive pressure environment in the area, at the same time, oxygen and carbon dioxide sensors are set up to monitor the data in the area, to create a low-oxygen environment under disaster conditions, and reference personnel enter the extreme condition simulation area to conduct blind wearing of self-rescuer under real extreme conditions.

[0024] The beneficial effects of the above technical solutions are that the mine disaster emergency escape training system provided by the present application deploys an identity recognition system at the entrance to accurately verify the identity of the students, ensures that the training records are accurately bound, creates a dangerous and urgent environment atmosphere that requires wearing a self-rescuer by simulating dangerous scenes such as mine fire, high temperature, thick smoke and collapse, and all mine personnel can carry out self-rescuer operation training in a tense state, detects the operation standard and completion time of the training personnel through AI video analysis technology, wears a smart watch during the training process to monitor the key physiological indicators of the students during the assessment process in real time, and assists in evaluating their psychological stress resistance and emergency adaptability. BRIEF DESCRIPTION OF DRAWINGS

[0025] Fig. 1 is a structural schematic diagram of a mine disaster emergency escape practical training system provided by an embodiment of the present application. DETAILED DESCRIPTION

[0026] The specific embodiments of the present application are described in further detail below in conjunction with the accompanying drawings and embodiments. The following embodiments are used to illustrate the present application, but are not used to limit the scope of the present application.

[0027] In this embodiment, the mine disaster emergency escape practical training system, as shown in Fig. 1, includes a roadway simulation module, a preparation area module, a self-rescuer blind wearing practical training module, a self-rescuer AI visual identification examination system, an examination confirmation module, and a self-rescuer training management system. Figure 1

[0028] The roadway simulation module is used for training and examination of emergency escape, and is used for simulating a roadway environment according to an actual roadway structure and simulating a roadway disaster environment.

[0029] The preparation area module is located in the simulated roadway, and is used for identification and registration of reference personnel information.

[0030] The self-rescuer blind wearing practical training module is located in the simulated roadway, and is used for wearing training of a self-rescuer by the reference personnel.

[0031] The examination confirmation module is located in the simulated roadway, and is used for examination of a result of wearing the self-rescuer by the reference personnel, returning the result and time to the self-rescuer training management system, confirming an examination time, and scoring according to a completion time.

[0032] In this embodiment, the preparation area module includes an identity recognition system and a health detection system; the identity recognition system includes iris recognition verification, face recognition, card registration, and an alcohol detection device; after the reference personnel enter the examination site preparation area, the iris recognition verification is performed first, and then the face recognition is performed, and then the card registration is performed, and then the alcohol detection is performed.

[0033] The health detection system includes a smart watch, and the physiological index data such as heart rate and blood oxygen saturation of the student in the simulation disaster environment in the practical operation examination process are monitored in real time through the smart watch, and the physiological reaction intensity of the student in the simulation disaster environment is analyzed, and the psychological stress resistance and emergency adaptability of the student are assisted to be evaluated; all physiological index data, student identity, operation video, and examination result are automatically bound and stored, and a complete training file is formed.

[0034] ​In the embodiment, the self-rescuer blind wearing practical training module is provided with a self-rescuer and a disaster environment simulation device; the disaster environment simulation device comprises a roof collapse simulation device, a flame simulation device, a sound atmosphere rendering system, a light atmosphere rendering system, a mine laser methane sensor, a carbon monoxide sensor, and a sound and light alarm; and the disaster atmosphere is created in cooperation. The roof collapse simulation device is used for simulating a roof fall accident scene in a coal mine; the flame simulation device is used for simulating a fire accident scene in a coal mine; the coal and gas outburst simulation device is used for simulating a coal and gas outburst scene in a coal mine, and the outburst simulation coal block is used for simulating the outburst of gas by spraying smoke by the smoke machine; the sound atmosphere rendering system is used for simulating the premonitory sound when the coal and gas outburst, such as the splitting sound, the thundering sound, the machine gun sound, the resonant coal gun, and the creaking sound when the gas passes through the water-containing fracture; the disaster scene sound of the outburst and the gas gushing when the coal and gas outburst is simulated; the burning and collapse sound when the fire occurs is simulated, and the sound of the roof collapse in the roof fall accident is simulated. The light atmosphere rendering system is used for simulating the light effect of the fire scene; the mine laser methane sensor and the carbon monoxide sensor are used for simulating the sensor alarm when the coal and gas outburst and the fire, and the sensor can be set in different ranges according to the disaster level to simulate the alarm by random numbers; the sound and light alarm is used for realizing the disaster evacuation alarm when the disaster occurs, and the disaster evacuation guide information is issued.

[0035] The mine intrinsic safety type infrared thermal imaging processing camera is further arranged in the self-rescuer blind wearing practical training module, the internal environment is monitored in real time, the video recognition of the low-visibility scene in the disaster scene is realized, and the self-rescuer wearing action of the training personnel is recognized; and the air leading and exhausting device is arranged in the self-rescuer blind wearing practical training module, and is used for exhausting the smoke in the area after the training is completed.

[0036] The AI camera is arranged in the examination and confirmation area module, the AI camera collects the process that the reference personnel wear the self-rescuer, and transmits the process to the self-rescuer AI visual recognition examination system; the self-rescuer AI visual recognition examination system confirms whether the examination personnel wear correctly, returns the examination result and the time to the self-rescuer training management system, confirms the examination time, and scores according to the completion time.

[0037] In the embodiment, the self-rescuer AI visual recognition examination system adopts a multi-task cooperative detection architecture, integrates three functions of human body target detection, number plate recognition, and self-rescuer wearing state detection, and realizes the intelligent binding of the personnel identity and the self-rescuer wearing state through the unique miner number plate;

[0038] The self-rescuer AI visual recognition assessment system adopts an improved yolov5 model for human target detection, number plate recognition and self-rescuer wearing state detection. The improved yolov5 model is based on the yolov5 model, and fuses a double-path enhancement mechanism. In the image preprocessing stage, an improved U-Net-based illumination perception subnetwork is deployed to make the image illumination uniform. The encoder of the improved U-Net-based illumination perception subnetwork adopts five layers of depth separable convolution, and the kernel sizes are 5*5, 5*5, 3*3, 3*3 and 3*3 in turn, so as to extract illumination features step by step. The decoder constructs an illumination distribution map through hole convolution and cross-layer connection.

[0039] In the feature extraction stage, a multi-scale smoke robust feature extraction module is designed. The multi-scale smoke robust feature extraction module is constructed based on depth separable convolution to form a four-layer pyramid structure with a step of two for downsampling. Each layer of features generates a channel weight vector through an SE attention module. Through an experimentally calibrated threshold function, the smoke-sensitive channels are suppressed, and the key features are preserved.

[0040] The multi-scale smoke robust feature extraction module is embedded into the backbone network of the yolov5 model to ensure the stability of detection.

[0041] The assessment confirmation area module is connected with the self-rescuer training management system. The results of scoring and the information captured by the AI camera can be displayed on the display screen.

[0042] In this embodiment, the simulated roadway further comprises a compressed air self-rescue practical training area module and an extreme condition simulation area module. The compressed air self-rescue practical training area is provided with a one-key pop-up compressed air self-rescue device. In this embodiment, a ZYJ-M6 compressed air self-rescue device is used. The device is designed in combination and has the functions of air supply and water supply. It also has the characteristics of pressure reduction, air volume adjustment, and noise reduction. The device is supplied with air and water from the ground, and is equipped with six compressed air self-rescue bags (i.e. breathing masks) and six drinking water hoses, which can be used by six people at the same time. When there is a situation that endangers people's lives, such as coal and gas outburst in the mine, the staff can open the box door of the device, open the pneumatic valve, and put on the mask to breathe until rescue. At this time, the compressed air self-rescue bag is automatically popped out by pressing the switch button, realizing rapid wearing and automatic air supply.

[0043] Meanwhile, the extreme condition simulation area module continuously injects carbon dioxide gas into the enclosed area of the simulated roadway through a carbon dioxide compression steel cylinder. A single exhaust system is provided in the area to maintain a positive pressure environment in the area. Oxygen and carbon dioxide sensors are provided in the area to monitor the data in the area, to create a low-oxygen environment under disaster conditions, and to enable personnel to enter the extreme condition simulation area for real extreme condition self-rescuer blind wearing practical training.

[0044] In this embodiment, the identity recognition system is deployed at the entrance of the simulated roadway to accurately verify the identity of the students, ensure the accurate binding of the training records, and set up an alcohol detection device to screen the drinking behavior of the students and eliminate the major safety hazards of drunk driving from the source;

[0045] In the simulated roadway, through dynamic simulation of dangerous scenes such as mine fire, high temperature, thick smoke, and collapse, an urgent and dangerous environment atmosphere is created that requires wearing a self-rescuer, so that all mine personnel can carry out self-rescuer operation training in a tense state, meet the 10-20 person single batch group training, and detect the reference personnel operation standardization and completion time through AI video analysis technology. The reference personnel wear smart watches throughout the training process, and the key physiological indicators (such as heart rate, blood oxygen saturation, etc.) of the reference personnel in the examination process are monitored in real time to assist in evaluating their psychological stress resistance and emergency adaptability. After the self-rescuer examination is completed, the students enter the compressed air self-rescue device operation area to complete the whole process operation training of the compressed air self-rescue device, and improve the emergency escape ability of all mine personnel and the safety management level of the enterprise.

[0046] In this embodiment, the main training process of the mine disaster emergency escape training system is as follows:

[0047] 1. After the reference personnel enter the preparation area of the examination site, the iris recognition is used to verify the identity, and then the digital number plate with RFID is swiped to register the number;

[0048] 2. The reference personnel leave the preparation area, and the self-rescuer training management system records the training start time; (the reference personnel enter in turn by swiping their faces, and a red and green light device is arranged in the area to indicate the reference personnel to pass in order).

[0049] 3. The reference personnel pass through the self-rescuer blind wearing training area, and the self-rescuer training management system controls the roof collapse simulation device, flame simulation device, sound atmosphere rendering system, and light atmosphere rendering system in the area to create a disaster environment. The mine-used laser methane sensor, carbon monoxide sensor, and sound and light alarm alarm to create a disaster atmosphere, and the mine-used intrinsically safe infrared thermal imaging processing camera in the self-rescuer blind wearing training area monitors in real time; (the system has an air guiding and exhausting device that can remove the smoke in the area after the training is completed).

[0050] 4. After passing through the self-rescuer blind wearing training area, the personnel enter the examination confirmation area, and after correctly wearing the self-rescuer, they raise their hands to confirm the completion of the examination. The AI camera confirms whether the examination personnel wear correctly through the self-rescuer AI visual identification examination system, returns the examination results and wearing time to the self-rescuer training management system, confirms the examination time, and scores according to the completion time. The examination confirmation area is equipped with a display screen to display the scores of the reference personnel.

[0051] In this embodiment, a monitoring display screen is installed outside the training area to display key information in real time, such as video pictures of reference personnel operating process, examination results, and analysis and statistics results, etc.

[0052] In this embodiment, an 8m extreme condition simulation area is provided outside the blind wearing training area of the self-rescuer. Carbon dioxide gas is continuously injected into the enclosed area in the simulated tunnel through a carbon dioxide compressed steel cylinder in the extreme condition simulation area. A single exhaust system (composed of pipelines and one-way exhaust valves) is provided in the area to maintain a positive pressure environment in the area. Oxygen sensors and carbon dioxide sensors are provided in the area to monitor the data in the area, to create a low-oxygen environment under disaster conditions. The reference personnel enter the extreme condition simulation area for blind wearing training of the self-rescuer under real extreme conditions. After passing through the extreme condition simulation area, the reference personnel enter the examination confirmation area. After correctly wearing the self-rescuer, the reference personnel raise their hands to confirm the completion of the examination. The AI camera confirms whether the reference personnel wear correctly through the self-rescuer AI visual recognition examination system. The examination results and time are returned to the self-rescuer training management system. The examination time is confirmed and scored according to the completion time. The examination confirmation area is equipped with a display screen to display the reference personnel's score.

[0053] In this embodiment, the software system involved in the mine disaster emergency escape training system is:

[0054] (1) Self-rescuer training management system: It realizes the unified scheduling of training resources and the centralized management of data, including student information management (input, grouping), student examination results management (input, storage, analysis), historical examination data statistics and query, system parameter configuration, user permission management, report generation and export, etc. It connects and coordinates the hardware system and other software systems.

[0055] (2) Self-rescuer AI visual recognition examination system: It uses cameras deployed in the simulated tunnel, combined with AI visual recognition technology, to capture the whole process video of the students wearing and operating the self-rescuer in the simulated disaster area in real time. Through deep learning algorithm, it automatically identifies each operation step of the students, accurately analyzes the standardization and completion time of the operation, automatically scores based on the preset standard operation process, generates the examination results, reduces the subjectivity and error of manual judgment, and improves the examination efficiency and accuracy.

[0056] In this embodiment, the hardware devices involved in the mine disaster emergency escape training system are:

[0057] (1) Simulated tunnel (training cabin): It simulates the mine tunnel environment and provides a relatively closed, controllable, and safe physical space for 10-20 people to form a group for immersive practical operation examination.

[0058] (2) Disaster environment simulation system: integrated in the simulated tunnel, using technical means to simulate the typical dangerous environment atmosphere when a mine disaster occurs, such as simulated smoke (non-toxic and harmless smoke), firelight (LED simulated light effect), high temperature (hot air blower), collapse (simulated collapse device), noise (explosion sound, etc.), aiming to stimulate the tension of the students and train their ability to operate the self-rescuer calmly under the pressure of the real environment.

[0059] (3) Self-rescuer: self-rescuer equipment or its simulation device for training, including pressure gauge, valve control, repeatable inflation / deflation device, detachable mouthpiece, etc. After identity recognition, each student receives a disposable and independent packaged mouthpiece. During training, the personal mouthpiece is inserted into the universal interface of the self-rescuer. After training, the student removes and discards the disposable mouthpiece, and puts the reusable self-rescuer into the designated recycling container for centralized professional disinfection.

[0060] (4) Identity recognition system: deployed at the entrance of the simulated tunnel, used for quick and accurate identification of the identity of the students entering the training cabin. The system identifies the student information and automatically associates the student's identity information with the subsequent training process video, AI assessment results, health data, etc. to ensure clear attribution of training data and traceable records. An alcohol detection device is deployed at the entrance of the simulated tunnel. Only after passing the alcohol content detection, the student can enter the practical operation examination. If the student's alcohol content exceeds the standard, the system will issue an alarm and prohibit entry into the examination area, and report to the administrator.

[0061] (5) Health monitoring system: by equipping with smart watches, real-time monitoring of key physiological indicators (such as heart rate, blood oxygen saturation, etc.) of students during the practical operation examination in the simulated disaster environment, analyzing the physiological reaction intensity of the reference personnel in the simulated disaster environment, assisting in evaluating their psychological stress resistance and emergency adaptability. All physiological indicator data are automatically bound and stored with the student's identity, operation video, and examination results to form a complete training file.

[0062] (6) Large screen display system: a monitoring display screen is installed outside the training cabin to display real-time key information such as video pictures of the reference personnel's operation process and examination results.

[0063] (7) Sound box system: used for playing realistic environmental sound effects to enhance the immersion of the disaster scene, and for system voice prompts and remote voice command and guidance of the management personnel to the student group or specific students.

[0064] (8) compressed air self-rescue operation device: a real compressed air self-rescue device is arranged in the simulated tunnel, the student presses the compressed air self-rescue device door lock switch, opens the box door, each set of compressed air self-rescue device contains 6 oxygen masks, presses the corresponding breathing mask switch, pops out the breathing mask, takes the light breathing mask and puts it on the mouth, automatically pulls out the safety switch to ventilate, and the student can train through the whole process of immersive operation to strengthen the emergency operation skills in disaster environment.

[0065] The mine disaster emergency escape practical training system of the present application has the following advantages:

[0066] 1. Whole-process training management: simulation training→intelligent examination→score display.

[0067] 2. Simulated disaster environment scene: simulate disaster environments such as fire, high-temperature smoke, and collapse to improve the psychological quality and practical operation ability of the trainees, and strengthen the muscle memory and emergency response ability.

[0068] 3. Standardized operation and accurate examination: real-time detection of 30s limited operation, AI action recognition technology automatically monitors the operation details of the students (such as “opening the cover, biting the mouthpiece, sealing the nose clip, etc.), and judges the operation integrity.

[0069] 4. Automatic scoring of examination results: with data collection, scoring, storage and query functions, it can collect operation data in real time as the basis for scoring and automatically score.

[0070] 5. High-concurrency group training capability: 10-20 people can be trained in groups, two people operate side by side, the next batch of students starts operation with a 5-10 meter interval, until all the students in the group are examined, and the system gives the examination situation of each student.

[0071] 6. Repeatability: self-rescuers that meet the requirements of the new standard can be set, and disposable mouthpieces are provided for each trainee each time.

[0072] 7. Health monitoring management: the heart rate, blood oxygen and other parameters of the trainees can be monitored in real time.

[0073] In this embodiment, the functions and system parameters to be realized by the mine disaster emergency escape practical training system are as follows:

[0074] 1. With AI visual recognition function, real-time capture of the whole process video of the students wearing and operating the self-rescuer in the simulated disaster area, accurate analysis of the standardization and completion time of the operation.

[0075] 2. With student (reference person) information management (input, grouping), student examination score management (input, storage, analysis), historical examination data statistics and query, system parameter configuration, user permission management, report generation and export, etc.

[0076] 3. Camera management: The system supports configuration management of camera information, including algorithm model binding, monitoring area binding, camera basic information binding, and streaming service binding functions. Binding cameras support custom monitoring area calibration functions.

[0077] 4. With an identity recognition system for quickly and accurately identifying the identity of trainees entering the training cabin and automatically associating the trainee's identity information with subsequent training process videos, AI examination results, health data, etc.

[0078] 5. With well entry alcohol detection function, after the alcohol content detection is qualified, it can enter the practical operation examination.

[0079] 6. The system supports group training and examination of 10-20 people.

[0080] 7. Support for compressed air self-rescue device training and exercise functions.

[0081] 8. With real-time monitoring of training personnel's heart rate, blood oxygen and other parameters.

[0082] 9. Technical parameters of mine intrinsically safe infrared thermal imaging processing camera:

[0083] (1) Explosion-proof type: Mine intrinsically safe type;

[0084] (2) Explosion-proof symbol: ExibIMb;

[0085] (3) Main technical indicators;

[0086] a. Image: color;

[0087] b. Minimum illumination: ≤0.05lx;

[0088] c. Horizontal resolution: ≥400 lines;

[0089] d. Gray scale: ≥7 levels;

[0090] (4) Ethernet electrical port;

[0091] a. Interface method: RJ45;

[0092] b. Interface quantity: 3-way;

[0093] c. Transmission rate: 100Mbps;

[0094] d. Maximum transmission distance: 50m;

[0095] (5) Ethernet optical port;

[0096] a. Interface method: SC single-mode double fiber;

[0097] b. Interface quantity: 2-way;

[0098] c. Light wavelength: 1310nm;

[0099] d. Transmission rate: 100Mbps;

[0100] e. Transmitting optical power: -20dBm~0dBm;

[0101] f. Maximum transmission distance: 10km.

[0102] (6) AI camera has the function of seeing the real object under the lowest illumination.

[0103] (7) AI camera has the function of converting the collected real-time image into Ethernet electrical signal and optical signal output.

[0104] (8) AI camera has infrared light supplement function.

[0105] 10. Self-rescuer AI vision recognition assessment system computing server technical parameters:

[0106] (1) Processor: integrated 2 Kunpeng 920 series (3210) processors (24Core, 2.6GHz);

[0107] (2) Memory: 4 memory 32GB DDR4 RECC;

[0108] (3) Hard disk: SSD 480GB SATA 6GB / s 2.5 read-intensive;

[0109] (4) Hard disk: hard disk HDD8T SATA enterprise class 3.57200;

[0110] (5) Array card: SR760-M(Avago3416) SAS / SATA RAID card-RAID0,1,10-12Gb / s-noCache card;

[0111] (6) Video analysis card: 1 video analysis card Atlas300VPro24GB DDR4X;

[0112] (7) Tray: hard disk rack 3.5 inch hard disk tray;

[0113] (8) Power supply: Kunpeng Ascend series general 900W;

[0114] 11. Self-rescuer training management system server technical parameters:

[0115] (1) CPU: 2 2.2G / 13.75M / 10C / 20T / 85W;

[0116] (2) Memory: 2 x 32GB DDR4 RECC;

[0117] (3) Hard disk: SSD 480GB SATA;

[0118] (4) Hard disk: HDD4T SATA enterprise class;

[0119] (5) Audio: with at least 1 interface;

[0120] (6) Power supply: 850W single;

[0121] (7) With USB interface, (including keyboard, mouse and 27-inch display);

[0122] (8) Operating system: Windows.

[0123] 12. Streaming media server;

[0124] (1) Provide stable webrtc, rtsp / rtmp, etc. Live streaming service, based on GPU hardware acceleration, realize high stability, high load, low resource consumption;

[0125] (2) The service also supports intelligent and stable bit rate, supports h264 industry common video encoding real-time conversion service;

[0126] (3) Support industry common hls and rtsp / rtmp protocol video intelligent real-time pull stream playing service, based on industry standard nginx underlying distribution architecture, provide high concurrency, low delay, second-level loading live pull stream solution;

[0127] (4) Based on the business system and intelligent video stream interception and compression protocol, support intelligent on-demand video playback service;

[0128] (5) Provide complete closed-loop solution to realize end-to-end live streaming, real-time playback and real-time recording, simple configuration and easy operation.

[0129] Finally, it should be noted that: the above examples are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing examples, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing examples, or make equivalent replacement for part or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope defined by the present application.

Claims

1. A mine disaster emergency escape training system, characterized in that: The application relates to a self-rescuer training system, which comprises a roadway simulation module, a preparation area module, a self-rescuer blind wearing practical training module, a self-rescuer AI visual identification examination system, an examination confirmation module and a self-rescuer training management system. The roadway simulation module is used for simulating a roadway disaster environment according to an actual roadway structure. The preparation area module is located in the simulated roadway and is used for identifying and registering personnel information. The self-rescuer blind wearing practical training module is located in the simulated roadway and is used for personnel to wear self-rescuers for practical training. The self-rescuer blind wearing practical training module is provided with a self-rescuer with a detachable mouthpiece. The self-rescuer blind wearing practical training module is provided with a mine intrinsic safety type infrared thermal imaging processing camera and an air guiding and exhausting device. The mine intrinsic safety type infrared thermal imaging processing camera is used for monitoring the internal environment of the simulated roadway in real time. The air guiding and exhausting device is used for exhausting smoke in the area after training. The examination confirmation module is located in the simulated roadway and is used for examining the result of the reference personnel wearing self-rescuers, returning the result and time to the self-rescuer training management system, confirming the examination time and scoring according to the completion time. The examination confirmation module is provided with an AI camera, which collects the process of the reference personnel wearing self-rescuers and transmits the process to the self-rescuer AI visual identification examination system. The self-rescuer AI visual identification examination system adopts an improved yolov5 model for human target detection, number plate identification and self-rescuer wearing state detection. The improved yolov5 model is based on the yolov5 model and fuses a double-path enhancement mechanism. In the image preprocessing stage, an improved U-Net-based illumination perception subnetwork is deployed to make the image illumination uniform. The encoder of the improved U-Net-based illumination perception subnetwork adopts five-layer depth separable convolution to extract illumination features step by step. In the feature extraction stage, a multi-scale smoke robust feature extraction module is designed. The multi-scale smoke robust feature extraction module is based on depth separable convolution and constructs a four-layer pyramid structure with a step of two for downsampling. Each layer of features generates a channel weight vector through an SE attention module, suppresses smoke-sensitive channels through an experiment-calibrated threshold function and retains key features. The multi-scale smoke robust feature extraction module is embedded in the backbone network of the yolov5 model to ensure the stability of detection.

2. The mine disaster emergency escape practical training system according to claim 1, characterized in that: The preparation area module comprises an identity recognition system and a health detection system, which are used for identity recognition and health detection of personnel information.

3. The mine disaster emergency escape practical training system according to claim 2, characterized in that: The identity recognition system comprises an iris recognition verification device, a face recognition device, a card swiping registration device and an alcohol detection device. After the personnel enter the preparation area of the training site, the iris recognition verification device and the face recognition device are used to verify the identity, then the card swiping registration device is used to register the number of the digital number card with RFID, and the alcohol detection device is used to detect alcohol.

4. The mine disaster emergency escape practical training system according to claim 3, characterized in that: The health detection system uses a smart watch to monitor the physiological index data of the personnel in the process of practical operation examination in the simulated roadway in real time, and analyzes the physiological reaction intensity of the personnel in the simulated roadway disaster environment, so as to assist in evaluating the psychological pressure resistance and emergency adaptability of the personnel. All physiological index data, student identity information, operation video and examination results are automatically bound and stored to form a complete training file.

5. The mine disaster emergency escape practical training system according to claim 1, characterized in that: The examination confirmation module is provided with a display screen connected with the self-rescuer training management system. The scoring results and the information shot by the AI camera are displayed on the display screen.

6. The mine disaster emergency escape practical training system according to claim 1, characterized in that: The simulated roadway is also provided with a compressed air self-rescue training area and an extreme condition simulation area. The compressed air self-rescue training area is used for developing compressed air self-rescue device training, and is provided with a one-key pop-up compressed air self-rescue device to realize rapid wearing and automatic air supply. The extreme condition simulation area continuously injects carbon dioxide gas into the closed area of the simulated roadway through a carbon dioxide compressed steel cylinder, and is provided with a single exhaust system to maintain the positive pressure environment in the area. At the same time, oxygen and carbon dioxide sensors are arranged to monitor the data in the area, so as to create a low-oxygen environment under disaster conditions, and the personnel enter the extreme condition simulation area to conduct blind wearing training of the self-rescuer under real extreme conditions.

Images