Mine emergency disaster avoidance channel detection method and system and storage medium

Abstract

The invention relates to the technical field of mine channel detection, in particular to a mine emergency disaster avoidance channel detection method and system and a storage medium, and the method comprises the steps: collecting the size feature data of each mine sub-channel, comparing the size feature data with preset standard size feature data, preliminarily analyzing whether the size of each mine sub-channel needs to be adjusted, and if not, determining whether the size of each mine sub-channel needs to be adjusted; if yes, collecting real-time environment characteristic data of each mine sub-channel, matching the real-time environment characteristic data with a preset safe environment interval, generating channel environment analysis data according to a matching result, and if the result is safe, ending the mine channel detection operation; if the mine sub-channel is unsafe, acquiring position data of the mine sub-channel with the environment abnormity, acquiring corresponding environment image feature data, inputting the data into a channel environment analysis model to accurately analyze a real-time environment abnormity type of the mine sub-channel with the environment abnormity, and finally generating mine channel detection data. And the information is pushed to the mine disaster-avoiding channel detection platform, so that intelligent detection of the mine emergency disaster-avoiding channel is realized.

Description

Technical Field

[0001] This invention relates to the field of mine passage detection technology, specifically to a method, system, and storage medium for detecting emergency escape passages in mines. Background Technology

[0002] In mining operations, emergency escape routes serve as the "lifeline" for miners' safety, and their safety and reliability are of paramount importance. However, due to factors such as complex geological conditions, long-term mining activities, and aging equipment, emergency escape routes often face numerous safety hazards.

[0003] Existing mine passage inspection technologies struggle to comprehensively inspect mine passages based on multiple pieces of mine passage data, resulting in poor quality of mine passage inspection operations. Furthermore, when monitoring real-time conditions within the mine, they often only analyze whether any anomalies have occurred in the mine passages, neglecting the types of anomalies within the mine passages, leading to incomplete inspection results. Summary of the Invention

[0004] To address the problems in related technologies, this invention provides a method, system, and storage medium for detecting emergency escape routes in mines, thereby overcoming the aforementioned technical problems existing in the prior art.

[0005] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a method for detecting emergency escape passages in mines, comprising the following steps: S1. Divide the mine emergency escape passage into several mine sub-passages and collect the dimensional characteristic data of each mine sub-passage in real time; S2. Analyze the size characteristics of each mine sub-channel based on the size characteristic data and the preset standard size characteristic data, and generate a mine sub-channel size adjustment instruction; If adjustments are required, data on abnormal dimensions will be collected and pushed to the mine emergency escape passage detection platform, and the passage dimensions will be adjusted accordingly. If no adjustment is needed, proceed directly to S3; S3. Collect real-time environmental characteristic data of each mine sub-channel; S4. Match the real-time environmental feature data with the preset safe environment range, and generate channel environment analysis data based on the matching results; If it is for safety reasons, then the current mine passage inspection operation shall be terminated. If it is unsafe, collect real-time environmental anomaly location data; S5. Real-time environmental image feature data corresponding to the real-time environmental anomaly location data is collected by an image capturing device to generate real-time environmental anomaly image feature data. S6. Based on the real-time environmental anomaly image feature data and the channel environment analysis model, perform environmental anomaly type analysis and processing to generate real-time environmental anomaly type text feature data. S7. Based on the real-time environmental anomaly location data, the real-time environmental anomaly image feature data, and the real-time environmental anomaly type text feature data, generate mine passage detection data and push it to the mine disaster avoidance passage detection platform.

[0006] Preferably, the specific steps for dividing the mine emergency escape passage into several mine sub-passages and collecting the dimensional characteristic data of each mine sub-passage in real time are as follows: S11. Set several channel division nodes, and divide the emergency escape channels of the target mine according to the channel division nodes to obtain the mine sub-channel set. ,in, Indicates the A mine tunnel This indicates the total number of mine tunnels; S12. Collect the dimensional feature data of each mine sub-channel in the mine sub-channel set using a 3D laser scanner to obtain the dimensional feature dataset. ,in, Indicates the Dimensional feature data corresponding to each mine sub-channel; The dimensional feature data includes width data, height data, and tilt angle data.

[0007] Preferably, the dimensional characteristics of each mine sub-channel are analyzed based on the dimensional characteristic data and the preset standard dimensional characteristic data to generate a mine sub-channel dimensional adjustment instruction. If adjustment is required, the abnormal dimensional location data is collected and pushed to the mine disaster avoidance channel detection platform, and the channel dimensional adjustment operation is executed. If no adjustment is required, the process proceeds directly to the following steps in S3: S21. Set standard size feature data, and compare each size feature data in the size feature dataset with the standard size feature data in turn; If all the size feature data in the size feature dataset are greater than or equal to the standard size feature data, it means that all mine sub-passes meet the passage standards. The output mine sub-pass size adjustment instruction is that no adjustment is needed, and the process directly enters S3. Otherwise, it indicates that a mine sub-passage does not meet the passage standards, and an adjustment instruction for the mine sub-passage size is output indicating that adjustment is needed. The location data of the mine sub-passage whose size feature data is smaller than the standard size feature data is then obtained to obtain a dataset of abnormal size locations. ,in, Indicates the Location data of mine sub-channels whose size feature data is smaller than the standard size feature data. This indicates the total number of data points indicating locations with abnormal dimensions. S22. The dataset of abnormal size locations is pushed to the mine emergency escape passage detection platform through the Internet of Things communication network, and the passage size adjustment operation is performed.

[0008] The dimensional feature data collected by the 3D laser scanner is accurately compared with the preset standard dimensional feature data. Based on the comparison results, it is initially analyzed whether each mine sub-passage meets the passage standards. The location data of mine sub-passages that do not meet the passage standards are pushed to the mine disaster avoidance passage detection platform to help staff quickly locate the mine sub-passages with dimensional abnormalities.

[0009] Preferably, the specific steps for collecting real-time environmental characteristic data of each mine sub-tunnel are as follows: S31. Real-time environmental feature data of each mine sub-channel is collected by different sensors to obtain a real-time environmental feature data matrix. as follows: , in, Indicates the number of data collected in real time by the sensor. The first mine tunnel Class-specific environmental feature data, The total number of categories representing environmental characteristic data; The various sensors include, but are not limited to, pressure sensors, level sensors, strain gauge sensors, and temperature sensors; The environmental characteristic data includes, but is not limited to, roof pressure data, water depth data, temperature data, and crack width data.

[0010] Preferably, the real-time environmental feature data is matched with a preset safe environment range, and channel environment analysis data is generated based on the matching result. If the range is safe, the current mine channel inspection operation ends; if the range is unsafe, the specific steps for collecting real-time environmental anomaly location data are as follows: S41. Utilize intelligent optimization algorithms to define safe environment intervals corresponding to various environmental characteristic data, generating a set of safe environment intervals. ,in, Indicates the The safe environment range corresponding to the class-specific environmental feature data; S411. Construct the environmental range search set, and set the current iteration number to . The maximum number of iterations is and search dimensions ; S412. Define the search range for the environmental intervals corresponding to various environmental feature data, and obtain the set of environmental interval search ranges as follows: ,in, and They represent the first The lower and upper limits of the search for the environmental range corresponding to the environmental feature data; Randomly generated within the search range set of the environmental interval The initial location set of the environmental interval search set is obtained by taking the environmental interval data as an example. ,in, Indicates the first element in the environmental range search set. The initial position for searching data within each environmental interval. This indicates the total number of environmental interval search data in the environmental interval search set; S413. Calculate the fitness value of each environment interval search data in the environment interval search set, sort the environment interval search data in the environment interval search set from largest to smallest fitness value, and select the environment interval search data with the highest fitness value as the current optimal solution; the fitness value calculation formula is as follows: , in, Indicates the first element in the environmental range search set. Fitness values ​​for searching data within a given environmental range. and These represent the misreporting weight and the missed report weight, respectively. Indicates the first element in the environmental range search set. The search data for the first environmental interval led to the first The probability of false positives being generated from security-related environmental characteristic data. Indicates the first element in the environmental range search set. The search data for the first environmental interval led to the first The probability that abnormal environmental feature data is missed. Indicates the correction value; S414. Update the balance factor; the updated formula is as follows: , in, This represents the balance factor in the current iteration process. This represents a random number that follows a uniform distribution between (0,1); like Then, each environmental interval search data in the environmental interval search set updates its position within the environmental interval search range set based on its own position and the position of a randomly selected environmental interval search data; the position update formula is as follows: , in, Indicates the first element in the environmental range search set. The environmental interval search data in the first The updated position within the dimensional environment range search range. and These respectively represent the first element in the environmental interval search set. The and the first The environmental interval search data in the first The current position within the dimensional environment range search range. and Both represent random numbers that follow a uniform distribution between (0,1); like Then, the position of each environmental interval search data in the environmental interval search set will be updated according to the position of the current optimal solution within the environmental interval search range set; the position update formula is as follows: , in, Indicates the first element in the environmental range search set. The location is updated after searching the environmental range data. and Both represent random numbers that follow a uniform distribution between (0,1). This indicates the position of the current best individual. and These respectively represent the first element in the environmental interval search set. The and the first The current position of the search data for each environmental interval. Indicates the control coefficient. Represents the Lévy flight function; S415. Update the variation rate of each environmental interval search data in the environmental interval search set. ; S416. Each environmental interval search data in the environmental interval search set undergoes a mutation operation based on the mutation rate within the environmental interval search range to update its position; the position update formula is as follows: , in, , and Both represent random numbers that follow a uniform distribution between (0,1). This indicates adaptive variable asynchronous length, and ,in, Indicates the initial variable asynchronous length. and These represent the minimum and maximum fitness values ​​of each environmental interval search data in the environmental interval search set, respectively. S417. Calculate the fitness value of each environmental interval search data in the environmental interval search set after position update. If the fitness value of environmental interval search data after position update is greater than the original fitness value, replace the original position with the new position of the environmental interval search data; otherwise, retain the original position. S418. Determine the current iteration number. Is it greater than or equal to the maximum number of iterations? If the current iteration number Greater than or equal to the maximum number of iterations If the fitness value is the highest, the search data for the environment interval with the highest fitness value is output, thus obtaining the set of safe environment intervals; otherwise, the current iteration number is... Increment by 1 and return S414; S42, Select the first element from the real-time environment feature data matrix. The real-time environmental feature data of each mine sub-channel is matched with the safe environment interval corresponding to the set of safe environment intervals; If the first in the real-time environment feature data matrix If all real-time environmental characteristic data within each mine sub-channel falls within the safe environment interval corresponding to the aforementioned safe environment interval set, then it indicates that the first... The mine tunnels are in a safe condition; Otherwise, it means the first One of the mine tunnels is in an unsafe condition; S43. Repeat the steps in S42 until all types of real-time environmental feature data in all mine sub-channels in the real-time environmental feature data matrix have been traversed. If all mine sub-passages are in a safe state, the output passage environment analysis data is safe, and it is pushed to the mine disaster avoidance passage detection platform through the Internet of Things communication network, thus ending this mine passage detection operation; If a mine sub-tunnel is found to be in an unsafe condition, the output tunnel environment analysis data will be marked as unsafe, and the location data of the unsafe mine sub-tunnel will be obtained to generate a real-time environmental anomaly location dataset. ,in, Indicates the Location data of a mine tunnel that is in an unsafe condition. This indicates the total number of real-time environmental anomaly location data.

[0011] By using intelligent optimization algorithms to accurately set safe environment ranges corresponding to various environmental feature data, and through multiple iterations of optimization, the safe environment ranges that most accurately identify environmental feature data are searched, which improves the accuracy of the search process and ensures its stability. At the same time, an adaptive variable asynchronous length is introduced during the algorithm execution process, which allows the environmental range search data to dynamically adjust the search range according to its own fitness value, thereby improving the search performance of the algorithm.

[0012] Preferably, the specific steps for generating real-time environmental anomaly image feature data by acquiring environmental image feature data corresponding to the real-time environmental anomaly location data in real time using an image capturing device are as follows: S51. Real-time acquisition of image feature data corresponding to the real-time environmental anomaly location data using a dual-spectrum camera device, generating a real-time environmental anomaly image feature dataset. ,in, Indicates the Image feature data corresponding to real-time environmental anomaly locations.

[0013] Preferably, the specific steps for generating real-time environmental anomaly type text feature data by performing environmental anomaly type analysis processing based on the real-time environmental anomaly image feature data and the channel environment analysis model are as follows: S61. By collecting image feature data of mine passages and corresponding text feature data of environmental anomaly types during various historical environmental anomalies through the mine disaster avoidance passage detection platform, a historical environmental anomaly image feature data matrix is ​​obtained. and historical environmental anomaly type text feature data matrix as follows: , , in, Indicates the occurrence of the first The first time the mine passage is in case of environmental anomalies Image feature data, Indicates the occurrence of the first The first time the mine passage is in case of environmental anomalies Text feature data corresponding to the environmental anomaly type of each image feature data; The textual feature data of the environmental anomaly types include, but are not limited to, minor water accumulation, severe water accumulation, minor roof cracks, moderate roof cracks, severe roof cracks, and excessively high temperatures; S62. Construct a channel environment analysis model based on the historical environmental anomaly image feature data matrix and the historical environmental anomaly type text feature data matrix; S621. Construct an initial convolutional neural network model and set the training error threshold and test accuracy threshold; S622. Set the training data ratio, and divide the historical environment anomaly image feature data matrix and the historical environment anomaly type text feature data matrix according to the training data ratio to obtain the historical environment anomaly image feature training data matrix, the historical environment anomaly type text feature training data matrix, the historical environment anomaly image feature test data matrix and the historical environment anomaly type text feature test data matrix. S623. Set the maximum number of training iterations, input the training data matrix of historical environment anomaly image features as training data and the training data matrix of historical environment anomaly type text features as training labels into the initial convolutional neural network model for training, and adjust the initial weights and initial biases of the initial convolutional neural network model according to the training results until the training error is less than the training error threshold or the number of training iterations is greater than the maximum number of training iterations, and then obtain the trained convolutional neural network model. S624. Input the historical environment anomaly image feature test data matrix as test data and the historical environment anomaly type text feature test data matrix as test labels into the trained convolutional neural network model for testing, calculate the accuracy of the test results. If the accuracy of the test results is greater than the test accuracy threshold, the channel environment analysis model is obtained; otherwise, return to S623 and retrain until the accuracy of the test results is greater than the test accuracy threshold. S63. Input each real-time environmental anomaly image feature data in the real-time environmental anomaly image feature dataset into the channel environment analysis model for analysis, and generate a real-time environmental anomaly type text feature dataset. ,in, Indicates the Text feature data of real-time environmental anomaly types corresponding to real-time environmental anomaly location data.

[0014] By matching the real-time environmental feature data with preset safe environment ranges, the system scientifically analyzes whether the real-time environment of each mine sub-channel is in a safe state. It also uses a dual-spectrum camera to collect image feature data of mine sub-channels in unsafe conditions in real time, and inputs this data into a pre-constructed channel environment analysis model for analysis. This accurately identifies the real-time environmental anomaly types of mine sub-channels in unsafe conditions, achieving intelligent detection of mine emergency escape channels. Furthermore, the initial model is trained and tested using historical environmental anomaly image feature data matrix and historical environmental anomaly type text feature data matrix to obtain the channel environment analysis model, providing a good and reliable tool for scientifically analyzing environmental anomalies occurring within mine sub-channels.

[0015] Preferably, the specific steps for generating mine passage detection data based on the real-time environmental anomaly location data, the real-time environmental anomaly image feature data, and the real-time environmental anomaly type text feature data, and pushing it to the mine disaster avoidance passage detection platform are as follows: S71. Combine the real-time environmental anomaly location dataset, the real-time environmental anomaly image feature dataset, and the real-time environmental anomaly type text feature dataset to generate mine tunnel detection data. ; S72. Transmit the mine passage detection data through the Internet of Things (IoT) communication network. The data is pushed to the mine emergency escape route monitoring platform and displayed on the screen.

[0016] The present invention also includes a mine emergency escape passage detection system, comprising a size feature data acquisition module, a size feature data discrimination module, a real-time environmental feature data acquisition module, a real-time environmental feature data analysis module, a real-time environmental anomaly image capture module, a real-time environmental anomaly type identification module, and a mine passage detection data construction module; The size feature data acquisition module sets several channel division nodes, and divides the emergency escape channel of the target mine according to the channel division nodes to obtain several mine sub-channels. The size feature data of each mine sub-channel is collected by a three-dimensional laser scanner. The size feature data discrimination module compares each size feature data in the size feature dataset with the standard size feature data in turn, and generates a mine sub-channel size adjustment instruction based on the comparison result. If adjustment is required, it collects the size anomaly location data and pushes it to the mine disaster avoidance channel detection platform; if adjustment is not required, it directly proceeds to S3. The real-time environmental feature data acquisition module collects real-time environmental feature data from various mine sub-channels using different sensors to obtain real-time environmental feature data. The real-time environmental feature data analysis module sets safe environmental ranges corresponding to various environmental feature data through intelligent optimization algorithms, matches the real-time environmental feature data with the safe environmental ranges, and generates channel environmental analysis data based on the matching results. If it is safe, the current mine channel detection operation ends; if it is unsafe, real-time abnormal location data is collected. The real-time environmental anomaly image capturing module acquires image feature data corresponding to the real-time environmental anomaly location data in real time through a dual-spectrum camera device, and generates real-time environmental anomaly image feature data. The real-time environmental anomaly type identification module collects image feature data of mine passages and corresponding environmental anomaly type text feature data online through the mine disaster avoidance passage detection platform when various environmental anomalies occur in history. Based on the image feature data and the environmental anomaly type text feature data, it constructs a passage environment analysis model and analyzes the real-time environmental anomaly image feature data to generate real-time environmental anomaly type text feature data. The mine passage detection data construction module combines the real-time environmental anomaly location dataset, the real-time environmental anomaly image feature dataset, and the real-time environmental anomaly type text feature dataset to generate mine passage detection data, which is then pushed to the mine disaster avoidance passage detection platform via the Internet of Things communication network and displayed on the screen.

[0017] By employing the above technical solution, the present invention provides a method, system, and storage medium for detecting emergency escape passages in mines, which has at least the following beneficial effects: 1. This invention collects dimensional feature data of each mine sub-channel and compares it with preset standard dimensional feature data to preliminarily analyze whether each mine sub-channel needs dimensional adjustment. At the same time, it collects environmental image feature data of mine sub-channels with environmental anomalies and inputs it into the channel environment analysis model to accurately analyze the real-time environmental anomaly type of the mine sub-channels with environmental anomalies. Finally, it generates mine channel detection data and pushes it to the mine emergency escape channel detection platform to realize intelligent detection of mine emergency escape channels.

[0018] 2. This invention accurately compares the dimensional feature data collected by a 3D laser scanner with the preset standard dimensional feature data. Based on the comparison results, it preliminarily analyzes whether each mine sub-passage meets the passage standards. The location data of mine sub-passages that do not meet the passage standards are pushed to the mine disaster avoidance passage detection platform to help staff quickly locate the mine sub-passages with dimensional abnormalities.

[0019] 3. This invention uses an intelligent optimization algorithm to accurately set the safe environment range corresponding to various environmental feature data. Through multiple iterations and optimizations, it searches for the safe environment range that best identifies the environmental feature data, thereby improving the accuracy and stability of the search process. At the same time, it introduces an adaptive variable asynchronous length during the algorithm execution process, which allows the environmental range search data to dynamically adjust the search range according to its own fitness value, thereby improving the search performance of the algorithm.

[0020] 4. This invention scientifically analyzes whether the real-time environment of each mine sub-channel is in a safe state by matching the real-time environmental feature data with a preset safe environment range. It also collects image feature data of mine sub-channels in an unsafe state in real time through a dual-spectrum camera device and inputs it into the channel environment analysis model for analysis. This accurately identifies the real-time environmental anomaly type of mine sub-channels in an unsafe state, realizes intelligent detection of mine emergency escape channels, and provides a good and reliable tool for scientifically analyzing environmental anomalies in mine sub-channels. Attached Figure Description

[0021] To more clearly illustrate the technical solutions of the embodiments of the invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the invention. For those skilled in the art, the drawings can be obtained from these drawings without creative effort.

[0022] Figure 1 A flowchart of the mine emergency escape passage detection method provided by the present invention; Figure 2 A schematic diagram of the modules of the mine emergency escape passage detection system provided by the present invention. Detailed Implementation

[0023] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0024] Example 1 is as follows: To address the limitations of existing mine passageway detection technologies in achieving comprehensive inspection and providing insufficient results, this embodiment proposes a method for detecting emergency escape passages in mines. This method collects dimensional feature data from various mine sub-passages and compares it with preset standard dimensional feature data to preliminarily analyze whether dimensional adjustments are needed for each sub-passage. Simultaneously, it collects environmental image feature data from mine sub-passages exhibiting abnormalities and inputs this data into a channel environment analysis model to accurately analyze the real-time environmental anomaly types, thus achieving intelligent detection of mine emergency escape passages. Figure 1 As shown, the method includes the following steps: S1. Divide the mine emergency escape passage into several mine sub-passages, and collect the dimensional characteristic data of each mine sub-passage in real time. This is a specific implementation plan of the method, and the detailed plan for this step is as follows: S11. Set several channel division nodes, and divide the emergency escape channels of the target mine according to the channel division nodes to obtain the mine sub-channel set. ,in, Indicates the A mine tunnel This indicates the total number of mine tunnels; S12. Collect the dimensional feature data of each mine sub-channel in the mine sub-channel set using a 3D laser scanner to obtain the dimensional feature dataset. ,in, Indicates the The dimensional characteristic data corresponding to each mine sub-tunnel includes width data, height data, and tilt angle data.

[0025] S2. Analyze the dimensional characteristics of each mine sub-tunnel based on the dimensional characteristic data and the preset standard dimensional characteristic data, and generate a mine sub-tunnel dimensional adjustment instruction. If adjustment is required, collect the dimensional anomaly location data and push it to the mine disaster avoidance passage detection platform to execute the passage dimensional adjustment operation; if no adjustment is required, proceed directly to S3. As a specific implementation plan of this method, the detailed plan for this step is as follows: S21. Set standard size feature data, and compare each size feature data in the size feature dataset with the standard size feature data in turn; If all size feature data in the size feature dataset are greater than or equal to the standard size feature data, it means that all mine sub-passes meet the passage standards. The output mine sub-pass size adjustment instruction is that no adjustment is needed, and the process directly enters S3. Otherwise, it indicates that a mine sub-passage does not meet the passage standards, and an adjustment instruction for the mine sub-passage size is output indicating that adjustment is needed. The location data of the mine sub-passage whose size feature data is smaller than the standard size feature data is then obtained, resulting in a dataset of abnormal size locations. ,in, Indicates the Location data of mine sub-tunnels whose size feature data is smaller than the standard size feature data. This indicates the total number of data points indicating locations with abnormal dimensions. S22. Push the dataset of abnormal size locations to the mine emergency escape passage detection platform through the Internet of Things communication network and perform passage size adjustment operations.

[0026] S3. Collect real-time environmental feature data for each mine sub-tunnel. S31. Collect real-time environmental feature data for each mine sub-tunnel using different sensors to obtain a real-time environmental feature data matrix. as follows: , in, Indicates the number of data collected in real time by the sensor. The first mine tunnel Class-specific environmental feature data, This represents the total number of categories of environmental characteristic data; different sensors include, but are not limited to, pressure sensors, level sensors, strain gauge sensors, and temperature sensors; environmental characteristic data includes, but is not limited to, roof pressure data, water depth data, temperature data, and crack width data.

[0027] S4. Match the real-time environmental feature data with the preset safe environment range, and generate channel environment analysis data based on the matching results. If the environment is safe, end the current mine channel detection operation; if it is unsafe, collect real-time abnormal location data as a specific implementation plan of this method. The detailed plan for this step is as follows: S41. Utilize intelligent optimization algorithms to define safe environment intervals corresponding to various environmental characteristic data, generating a set of safe environment intervals. ,in, Indicates the The safe environment range corresponding to the class-specific environmental feature data; S411. Construct the environmental range search set, and set the current iteration number to . The maximum number of iterations is and search dimensions ; S412. Define the search range for the environmental intervals corresponding to various environmental feature data, and obtain the set of environmental interval search ranges as follows: ,in, and They represent the first The lower and upper limits of the search for the environmental range corresponding to the environmental feature data; Randomly generated within the search range of the environmental interval. The initial location set of the environmental interval search set is obtained by taking the environmental interval data as an example. ,in, Indicates the first element in the environmental range search set. The initial position for searching data within each environmental interval. This indicates the total number of environment range search data in the environment range search set; S413. Calculate the fitness value of each environment interval search data in the environment interval search set. Sort each environment interval search data in the environment interval search set according to its fitness value from largest to smallest, and select the environment interval search data with the highest fitness value as the current optimal solution. The fitness value calculation formula is as follows: , in, Indicates the first element in the environmental range search set. Fitness values ​​for searching data within a given environmental range. and These represent the misreporting weight and the missed report weight, respectively. Indicates the first element in the environmental range search set. The search data for the first environmental interval led to the first The probability of false positives being generated from security-related environmental characteristic data. Indicates the first element in the environmental range search set. The search data for the first environmental interval led to the first The probability that abnormal environmental feature data is missed. Indicates the correction value; S414. Update the balance factor; the updated formula is as follows: , in, This represents the balance factor in the current iteration process. This represents a random number that follows a uniform distribution between (0,1); like Then, each environmental interval search data in the environmental interval search set updates its position within the environmental interval search range set based on its own position and the position of a randomly selected environmental interval search data; the position update formula is as follows: , in, Indicates the first element in the environmental range search set. The environmental interval search data in the first The updated position within the dimensional environment range search range. and These represent the first element in the environmental range search set. The and the first The environmental interval search data in the first The current position within the dimensional environment range search range. and Both represent random numbers that follow a uniform distribution between (0,1); like Then, the position of each environment interval search data in the environment interval search set will be updated according to the position of the current optimal solution; the position update formula is as follows: , in, Indicates the first element in the environmental range search set. The location is updated after searching the environmental range data. and Both represent random numbers that follow a uniform distribution between (0,1). This indicates the position of the current best individual. and These represent the first element in the environmental range search set. The and the first The current position of the search data for each environmental interval. Indicates the control coefficient. Represents the Lévy flight function; S415. Update the mutation rate of each environment range search data in the environment range search set. ; S416. The search data for each environmental interval in the environmental interval search set are updated in position by performing a mutation operation within the environmental interval search range set according to the mutation rate; the position update formula is as follows: , in, , and Both represent random numbers that follow a uniform distribution between (0,1). This indicates adaptive variable asynchronous length, and ,in, Indicates the initial variable asynchronous length. and These represent the minimum and maximum fitness values ​​of the search data in each environment interval search set, respectively. S417. Calculate the fitness value of each environment interval search data in the environment interval search set after position update. If the fitness value of the environment interval search data after position update is greater than the original fitness value, replace the original position with the new position of the environment interval search data; otherwise, retain the original position. S418. Determine the current iteration number. Is it greater than or equal to the maximum number of iterations? If the current iteration number Greater than or equal to the maximum number of iterations If the fitness value is the highest, the search data for the environment interval with the highest fitness value is output, thus obtaining the set of safe environment intervals; otherwise, the current iteration number is... Increment by 1 and return S414; S42. Select the first element from the real-time environment feature data matrix. The real-time environmental characteristic data of each mine sub-channel is matched with the corresponding safe environment intervals in the safe environment interval set. If the real-time environment feature data matrix contains the first... If all real-time environmental characteristic data within each mine sub-channel falls within the safe environment interval corresponding to the set of safe environment intervals, then it indicates that the first... The mine tunnels are in a safe condition; Otherwise, it means the first One of the mine tunnels is in an unsafe condition; S43. Repeat the steps in S42 until all types of real-time environmental feature data in all mine sub-channels in the real-time environmental feature data matrix have been traversed. If all mine sub-passages are in a safe state, the output passage environment analysis data is safe, and it is pushed to the mine disaster avoidance passage detection platform through the Internet of Things communication network, thus ending this mine passage detection operation; If a mine sub-tunnel is found to be in an unsafe condition, the output tunnel environment analysis data will be marked as unsafe, and the location data of the unsafe mine sub-tunnel will be obtained to generate a real-time environmental anomaly location dataset. ,in, Indicates the Location data of a mine tunnel that is in an unsafe condition. This indicates the total number of real-time environmental anomaly location data.

[0028] S5. Real-time acquisition of environmental image feature data corresponding to real-time environmental anomaly location data using an image capturing device, generating real-time environmental anomaly image feature data. S51. Real-time acquisition of image feature data corresponding to real-time environmental anomaly location data using a dual-spectrum camera device, generating a real-time environmental anomaly image feature dataset. ,in, Indicates the Image feature data corresponding to real-time environmental anomaly locations.

[0029] S6. Based on the real-time environmental anomaly image feature data and the channel environment analysis model, perform environmental anomaly type analysis and processing to generate real-time environmental anomaly type text feature data. As a specific implementation plan of this method, the detailed plan for this step is as follows: S61. By collecting image feature data of mine passages and corresponding text feature data of environmental anomaly types during various historical environmental anomalies through the mine disaster avoidance passage detection platform, a historical environmental anomaly image feature data matrix is ​​obtained. and historical environmental anomaly type text feature data matrix as follows: , , in, Indicates the occurrence of the first The first time the mine passage is in case of environmental anomalies Image feature data, Indicates the occurrence of the first The first time the mine passage is in case of environmental anomalies Text feature data corresponding to the environmental anomaly type of each image feature data; Environmental anomaly type text feature data includes, but is not limited to, minor water accumulation, severe water accumulation, minor roof cracks, moderate roof cracks, severe roof cracks, and excessively high temperatures; S62. Construct a channel environment analysis model based on the historical environmental anomaly image feature data matrix and the historical environmental anomaly type text feature data matrix; S621. Construct an initial convolutional neural network model and set the training error threshold and test accuracy threshold; S622. Set the training data ratio, and divide the historical environment anomaly image feature data matrix and the historical environment anomaly type text feature data matrix according to the training data ratio to obtain the historical environment anomaly image feature training data matrix, the historical environment anomaly type text feature training data matrix, the historical environment anomaly image feature test data matrix, and the historical environment anomaly type text feature test data matrix. S623. Set the maximum number of training iterations. Input the training data matrix of historical environmental anomaly image features and the training data matrix of historical environmental anomaly type text features as training labels into the initial convolutional neural network model for training. Adjust the initial weights and initial biases of the initial convolutional neural network model according to the training results until the training error is less than the training error threshold or the number of training iterations is greater than the maximum number of training iterations, and then obtain the trained convolutional neural network model. S624. Input the historical environment anomaly image feature test data matrix as test data and the historical environment anomaly type text feature test data matrix as test labels into the trained convolutional neural network model for testing. Calculate the accuracy of the test results. If the accuracy of the test results is greater than the test accuracy threshold, the channel environment analysis model is obtained; otherwise, return to S623 and retrain until the accuracy of the test results is greater than the test accuracy threshold. S63. Input the real-time environmental anomaly image feature data from the real-time environmental anomaly image feature dataset into the channel environment analysis model for analysis, and generate a real-time environmental anomaly type text feature dataset. ,in, Indicates the Text feature data of real-time environmental anomaly types corresponding to real-time environmental anomaly location data.

[0030] S7. Based on real-time environmental anomaly location data, real-time environmental anomaly image feature data, and real-time environmental anomaly type text feature data, mine passage detection data is generated and pushed to the mine disaster avoidance passage detection platform. As a specific implementation plan of this method, the detailed plan for this step is as follows: S71. Combine the real-time environmental anomaly location dataset, the real-time environmental anomaly image feature dataset, and the real-time environmental anomaly type text feature dataset to generate mine tunnel detection data. ; S72. Transmit mine passage detection data via IoT communication network The data is pushed to the mine emergency escape route monitoring platform and displayed on the screen.

[0031] Example 2 is as follows: See also Figure 2 A mine emergency escape passage detection system includes a size feature data acquisition module, a size feature data discrimination module, a real-time environmental feature data acquisition module, a real-time environmental feature data analysis module, a real-time environmental anomaly image capture module, a real-time environmental anomaly type identification module, and a mine passage detection data construction module. The size feature data acquisition module sets several channel division nodes, divides the emergency escape channels of the target mine according to the channel division nodes, and obtains several mine sub-channels. The size feature data of each mine sub-channel is collected by a 3D laser scanner. The size feature data discrimination module compares each size feature data in the size feature dataset with the standard size feature data in turn. Based on the comparison results, it generates a mine sub-channel size adjustment instruction. If adjustment is required, it collects the size anomaly location data and pushes it to the mine disaster avoidance channel detection platform; if adjustment is not required, it directly enters S3. The real-time environmental feature data acquisition module collects real-time environmental feature data from various mine sub-channels using different sensors to obtain real-time environmental feature data. The real-time environmental feature data analysis module uses intelligent optimization algorithms to set safe environmental ranges corresponding to various environmental feature data, matches the real-time environmental feature data with the safe environmental ranges, and generates channel environmental analysis data based on the matching results. If it is safe, the current mine channel detection operation ends; if it is unsafe, real-time abnormal location data is collected. The real-time environmental anomaly image capture module uses a dual-spectrum camera to collect image feature data corresponding to the real-time environmental anomaly location data and generate real-time environmental anomaly image feature data. The real-time environmental anomaly type identification module collects image feature data of mine passages and corresponding text feature data of environmental anomalies in history through the mine disaster avoidance passage detection platform. Based on the image feature data and the text feature data of environmental anomalies, it constructs a passage environment analysis model and analyzes the real-time environmental anomaly image feature data to generate real-time environmental anomaly type text feature data. The mine passage detection data construction module combines real-time environmental anomaly location dataset, real-time environmental anomaly image feature dataset, and real-time environmental anomaly type text feature dataset to generate mine passage detection data. This data is then pushed to the mine disaster avoidance passage detection platform via the Internet of Things communication network and displayed on the screen.

[0032] Those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing related hardware. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, this application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0033] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0034] The preferred embodiments of the invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention.

Claims

1. A method for detecting emergency escape routes in mines, characterized in that, The steps include: S1. Divide the mine emergency escape passage into several mine sub-passages and collect the dimensional characteristic data of each mine sub-passage in real time; S2. Analyze the size characteristics of each mine sub-channel based on the size characteristic data and the preset standard size characteristic data, and generate a mine sub-channel size adjustment instruction; If adjustments are required, data on abnormal dimensions will be collected and pushed to the mine emergency escape passage detection platform, and the passage dimensions will be adjusted accordingly. If no adjustment is needed, proceed directly to S3; S3. Collect real-time environmental characteristic data of each mine sub-channel; S4. Match the real-time environmental feature data with the preset safe environment range, and generate channel environment analysis data based on the matching results; If it is for safety reasons, then the current mine passage inspection operation shall be terminated. If it is unsafe, collect real-time environmental anomaly location data; S5. Real-time environmental image feature data corresponding to the real-time environmental anomaly location data is collected by an image capturing device to generate real-time environmental anomaly image feature data. S6. Based on the real-time environmental anomaly image feature data and the channel environment analysis model, perform environmental anomaly type analysis and processing to generate real-time environmental anomaly type text feature data. S7. Based on the real-time environmental anomaly location data, the real-time environmental anomaly image feature data, and the real-time environmental anomaly type text feature data, generate mine passage detection data and push it to the mine disaster avoidance passage detection platform.

2. The method for detecting emergency escape passages in mines according to claim 1, characterized in that, S1 includes the following steps: S11. Set several channel division nodes, and divide the emergency escape channels of the target mine according to the channel division nodes to obtain the mine sub-channel set. ,in, Indicates the A mine tunnel This indicates the total number of mine tunnels; S12. Collect the dimensional feature data of each mine sub-channel in the mine sub-channel set using a 3D laser scanner to obtain the dimensional feature dataset. ,in, Indicates the Dimensional feature data corresponding to each mine sub-channel; The dimensional feature data includes width data, height data, and tilt angle data.

3. The method for detecting emergency escape passages in mines according to claim 2, characterized in that, S2 includes the following steps: S21. Set standard size feature data, and compare each size feature data in the size feature dataset with the standard size feature data in turn; If all the size feature data in the size feature dataset are greater than or equal to the standard size feature data, it means that all mine sub-passes meet the passage standards. The output mine sub-pass size adjustment instruction is that no adjustment is needed, and the process directly enters S3. Otherwise, it indicates that a mine sub-passage does not meet the passage standards, and an adjustment instruction for the mine sub-passage size is output indicating that adjustment is needed. The location data of the mine sub-passage whose size feature data is smaller than the standard size feature data is then obtained to obtain a dataset of abnormal size locations. ,in, Indicates the Location data of mine sub-channels whose size feature data is smaller than the standard size feature data. This indicates the total number of data points indicating locations with abnormal dimensions. S22. The dataset of abnormal size locations is pushed to the mine emergency escape passage detection platform through the Internet of Things communication network, and the passage size adjustment operation is performed.

4. The method for detecting emergency escape passages in mines according to claim 3, characterized in that, The process of collecting real-time environmental feature data for each mine sub-channel includes the following steps: Real-time environmental feature data of each mine sub-tunnel was collected by different sensors to obtain a real-time environmental feature data matrix. as follows: ， in, Indicates the number of data collected in real time by the sensor. The first mine tunnel Class-specific environmental feature data, The total number of categories representing environmental characteristic data; The various sensors include, but are not limited to, pressure sensors, level sensors, strain gauge sensors, and temperature sensors; The environmental characteristic data includes, but is not limited to, roof pressure data, water depth data, temperature data, and crack width data.

5. The method for detecting emergency escape passages in mines according to claim 4, characterized in that, S4 includes the following steps: S41. Utilize intelligent optimization algorithms to define safe environment intervals corresponding to various environmental characteristic data, generating a set of safe environment intervals. ,in, Indicates the The safe environment range corresponding to the class-specific environmental feature data; S411. Construct the environmental range search set, and set the current iteration number to . The maximum number of iterations is and search dimensions ; S412. Define the search range for the environmental intervals corresponding to various environmental feature data, and obtain the set of environmental interval search ranges as follows: ,in, and They represent the first The lower and upper limits of the search for the environmental range corresponding to the environmental feature data; Randomly generated within the search range set of the environmental interval The initial location set of the environmental interval search set is obtained by taking the environmental interval data as an example. ,in, Indicates the first element in the environmental range search set. The initial position for searching data within each environmental interval. This indicates the total number of environmental interval search data in the environmental interval search set; S413. Calculate the fitness value of each environment interval search data in the environment interval search set, sort the environment interval search data in the environment interval search set from largest to smallest fitness value, and select the environment interval search data with the highest fitness value as the current optimal solution; the fitness value calculation formula is as follows: ， in, Indicates the first element in the environmental range search set. Fitness values ​​for searching data within a given environmental range. and These represent the misreporting weight and the missed report weight, respectively. Indicates the first element in the environmental range search set. The search data for the first environmental interval led to the first The probability of false positives being generated from security-related environmental characteristic data. Indicates the first element in the environmental range search set. The search data for the first environmental interval led to the first The probability that abnormal environmental feature data is missed. Indicates the correction value; S414. Update the balance factor; the updated formula is as follows: ， in, This represents the balance factor in the current iteration process. This represents a random number that follows a uniform distribution between (0,1); like Then, each environmental interval search data in the environmental interval search set updates its position within the environmental interval search range set based on its own position and the position of a randomly selected environmental interval search data; the position update formula is as follows: ， in, Indicates the first element in the environmental range search set. The environmental interval search data in the first The updated position within the dimensional environment range search range. and These respectively represent the first element in the environmental interval search set. The and the first The environmental interval search data in the first The current position within the dimensional environment range search range. and Both represent random numbers that follow a uniform distribution between (0,1); like Then, the position of each environmental interval search data in the environmental interval search set will be updated according to the position of the current optimal solution within the environmental interval search range set; the position update formula is as follows: ， in, Indicates the first element in the environmental range search set. The location is updated after searching the environmental range data. and Both represent random numbers that follow a uniform distribution between (0,1). This indicates the position of the current best individual. and These respectively represent the first element in the environmental interval search set. The and the first The current position of the search data for each environmental interval. Indicates the control coefficient. Represents the Lévy flight function; S415. Update the variation rate of each environmental interval search data in the environmental interval search set. ; S416. Each environmental interval search data in the environmental interval search set undergoes a mutation operation based on the mutation rate within the environmental interval search range to update its position; the position update formula is as follows: ， in, , and Both represent random numbers that follow a uniform distribution between (0,1). This indicates adaptive variable asynchronous length, and ,in, Indicates the initial variable asynchronous length. and These represent the minimum and maximum fitness values ​​of each environmental interval search data in the environmental interval search set, respectively. S417. Calculate the fitness value of each environmental interval search data in the environmental interval search set after position update. If the fitness value of environmental interval search data after position update is greater than the original fitness value, replace the original position with the new position of the environmental interval search data; otherwise, retain the original position. S418. Determine the current iteration number. Is it greater than or equal to the maximum number of iterations? If the current iteration number Greater than or equal to the maximum number of iterations If the fitness value is the highest, the search data for the environment interval with the highest fitness value is output, thus obtaining the set of safe environment intervals; otherwise, the current iteration number is... Increment by 1 and return S414; S42, Select the first element from the real-time environment feature data matrix. The real-time environmental feature data of each mine sub-channel is matched with the safe environment interval corresponding to the set of safe environment intervals; If the first in the real-time environment feature data matrix If all real-time environmental characteristic data within each mine sub-channel falls within the safe environment interval corresponding to the aforementioned safe environment interval set, then it indicates that the first... The mine tunnels are in a safe condition; Otherwise, it means the first One of the mine tunnels is in an unsafe condition; S43. Repeat the steps in S42 until all types of real-time environmental feature data in all mine sub-channels in the real-time environmental feature data matrix have been traversed. If all mine sub-passages are in a safe state, the output passage environment analysis data is safe, and it is pushed to the mine disaster avoidance passage detection platform through the Internet of Things communication network, thus ending this mine passage detection operation; If a mine sub-tunnel is found to be in an unsafe condition, the output tunnel environment analysis data will be marked as unsafe, and the location data of the unsafe mine sub-tunnel will be obtained to generate a real-time environmental anomaly location dataset. ,in, Indicates the Location data of a mine tunnel that is in an unsafe condition. This indicates the total number of real-time environmental anomaly location data.

6. The method for detecting emergency escape passages in mines according to claim 5, characterized in that, The generated real-time environmental anomaly image feature data includes: Image feature data corresponding to the real-time environmental anomaly location data are acquired in real time using a dual-spectrum camera device, generating a real-time environmental anomaly image feature dataset. ,in, Indicates the Image feature data corresponding to real-time environmental anomaly locations.

7. The method for detecting emergency escape passages in mines according to claim 6, characterized in that, S6 includes the following steps: S61. By collecting image feature data of mine passages and corresponding text feature data of environmental anomaly types during various historical environmental anomalies through the mine disaster avoidance passage detection platform, a historical environmental anomaly image feature data matrix is ​​obtained. and historical environmental anomaly type text feature data matrix ,as follows: ， ， in, Indicates the occurrence of the first The first time the mine passage is in case of environmental anomalies Image feature data, Indicates the occurrence of the first The first time the mine passage is in case of environmental anomalies Text feature data corresponding to the environmental anomaly type of each image feature data; The textual feature data of the environmental anomaly types include, but are not limited to, minor water accumulation, severe water accumulation, minor roof cracks, moderate roof cracks, severe roof cracks, and excessively high temperatures; S62. Construct a channel environment analysis model based on the historical environmental anomaly image feature data matrix and the historical environmental anomaly type text feature data matrix; S621. Construct an initial convolutional neural network model and set the training error threshold and test accuracy threshold; S622. Set the training data ratio, and divide the historical environment anomaly image feature data matrix and the historical environment anomaly type text feature data matrix according to the training data ratio to obtain the historical environment anomaly image feature training data matrix, the historical environment anomaly type text feature training data matrix, the historical environment anomaly image feature test data matrix and the historical environment anomaly type text feature test data matrix. S623. Set the maximum number of training iterations, input the training data matrix of historical environment anomaly image features as training data and the training data matrix of historical environment anomaly type text features as training labels into the initial convolutional neural network model for training, and adjust the initial weights and initial biases of the initial convolutional neural network model according to the training results until the training error is less than the training error threshold or the number of training iterations is greater than the maximum number of training iterations, and then obtain the trained convolutional neural network model. S624. Input the historical environment anomaly image feature test data matrix as test data and the historical environment anomaly type text feature test data matrix as test labels into the trained convolutional neural network model for testing, calculate the accuracy of the test results. If the accuracy of the test results is greater than the test accuracy threshold, the channel environment analysis model is obtained; otherwise, return to S623 and retrain until the accuracy of the test results is greater than the test accuracy threshold. S63. Input each real-time environmental anomaly image feature data in the real-time environmental anomaly image feature dataset into the channel environment analysis model for analysis, and generate a real-time environmental anomaly type text feature dataset. ,in, Indicates the Text feature data of real-time environmental anomaly types corresponding to real-time environmental anomaly location data.

8. The method for detecting emergency escape passages in mines according to claim 7, characterized in that, S7 includes the following steps: The real-time environmental anomaly location dataset, the real-time environmental anomaly image feature dataset, and the real-time environmental anomaly type text feature dataset are combined to generate mine tunnel detection data. ; The mine passage detection data is transmitted via an Internet of Things (IoT) communication network. The data is pushed to the mine emergency escape route monitoring platform and displayed on the screen.

9. A system for implementing the mine emergency escape passage detection method according to any one of claims 1-8.

10. A computer storage medium storing computer program instructions thereon, characterized in that, When the computer program instructions are executed by the processor, they implement the steps of the mine emergency escape passage detection method according to any one of claims 1 to 8.

Images