Intelligent mine monitoring system based on AI video intelligent assistance

Abstract

The invention discloses a mine intelligent monitoring system based on AI video intelligent assistance, and relates to the technical field of mine monitoring. Comprising an image data acquisition module, an image data preprocessing module, a geological feature extraction module, a machine learning intelligent evaluation module, a coal seam classification decision module, a stable coal seam operation module and a mixed coal seam speed regulation protection module, and the image data acquisition module acquires coal seam image data in the advancing direction of a mining machine in real time through a video monitoring system. Through an AI video intelligent monitoring method, coal seam images are collected in real time, geological changes are intelligently evaluated in combination with machine learning, and coal seam dynamic classification is achieved. In the mixed coal seam, the speed of a mining machine is reduced, and tool abrasion and fault risks are reduced; in the stable coal seam, high-efficiency operation at the preset speed is kept. Equipment protection and production efficiency are balanced through intelligent regulation and control, and economic benefits and stability of mining operation are improved.

Description

Technical Field

[0001] The present invention relates to the field of mine monitoring technology, and in particular to an AI video intelligently assisted mine intelligent monitoring system. Background Art

[0002] Intelligent mine monitoring based on AI video intelligence assistance is an intelligent management solution that combines artificial intelligence (AI) technology with a video monitoring system to monitor the mine environment, equipment operation and personnel activities in real time. Through cameras installed at key locations in the mine, the video monitoring system collects video data in real time and uses AI algorithms to analyze and process the video content. AI technology can identify potential risks in mines, such as equipment failures, environmental abnormalities (such as fire, smoke, landslides, etc.), and personnel behaviors that do not comply with safety regulations. At the same time, the system can automatically generate alarm information and immediately notify management personnel through technologies such as image recognition, behavior analysis, and target tracking, and even automatically take early warning measures according to set rules, thereby improving the safety, operational efficiency and intelligence level of the mine. This intelligent monitoring system not only improves the monitoring capabilities of mines, but also greatly reduces human operating errors, helping to achieve remote, automated, and refined mine management.

[0003] The prior art has the following deficiencies:

[0004] In the process of coal mining, mining machines usually operate at a fixed speed, which can maintain efficient and stable mining progress when the coal seam is uniform and the coal quality is pure. However, coal mine layers are often a mixture of coal, ore, rock and other minerals, and the hardness of some ores is significantly higher than that of coal (such as hard minerals such as quartz, feldspar, and iron ore). When the mining machine encounters these hard minerals, continuing to operate at a fixed speed may cause a series of serious problems. First, the high hardness of hard minerals will increase the friction between the tool and the ore layer, causing the tool to wear faster, thereby increasing the replacement frequency and maintenance costs. Secondly, some hard minerals (such as iron ore, quartzite, etc.) are extremely hard. When the tool is subjected to excessive friction and impact, it may crack or break, which will affect the normal operation of the equipment and even cause downtime and equipment failure. Finally, the presence of hard minerals will significantly reduce the cutting efficiency, and more force and time are required to handle these hard minerals, resulting in reduced operating efficiency and hindered mining progress.

[0005] The above information disclosed in this Background section is only for enhancement of understanding of the background of the present disclosure and therefore it may contain information that does not constitute the prior art that is already known to one of ordinary skill in the art. Summary of the invention

[0006] The purpose of the present invention is to provide an intelligent mine monitoring system based on AI video intelligent assistance. Through the AI ​​video intelligent assistance mine intelligent monitoring method, coal seam image data is collected in real time and combined with machine learning models to intelligently evaluate coal seam geological changes, so as to realize dynamic classification of coal seams. In mixed coal seams, the system reduces the forward speed of the mining machine, reduces the friction impact between the tool and the hard mineral, reduces the wear rate and failure risk, extends the tool life, and reduces downtime maintenance costs; in stable coal seams, the mining machine maintains a preset speed to ensure efficient operation. Through intelligent dynamic regulation, the solution achieves a balance between equipment protection and production efficiency, improves the economic benefits and stability of mining operations, and solves the problems in the above-mentioned background technology.

[0007] In order to achieve the above-mentioned purpose, the present invention provides the following technical solutions: an intelligent monitoring system for mines based on AI video intelligence assistance, including an image data acquisition module, an image data preprocessing module, a geological feature extraction module, a machine learning intelligent evaluation module, a coal seam classification decision module, a stable coal seam operation module and a mixed coal seam speed regulation protection module:

[0008] The image data acquisition module collects the coal seam image data in the direction of the mining machine in real time through the video monitoring system;

[0009] The image data preprocessing module preprocesses the acquired coal seam images to ensure that the acquired coal seam images are clear and accurately reflect the geological characteristics of the coal seams;

[0010] The geological feature extraction module extracts key features reflecting the geological changes of coal seams from the preprocessed image data after preprocessing the acquired image data, and analyzes the extracted geological change features under the monitoring window to quantify the distribution and changes of geological features;

[0011] The machine learning intelligent evaluation module inputs the analyzed features into a pre-trained machine learning model and intelligently evaluates the geological changes of the coal seam through the feature quantification data;

[0012] The coal seam classification decision module classifies coal seam geological changes into stable coal seams and mixed coal seams based on the evaluation results of the machine learning model;

[0013] Stable coal seam operation module: For stable coal seams, the ore layer is uniform and the coal quality is relatively pure. The mining machine mines at the preset forward speed to maintain mining efficiency;

[0014] The mixed coal seam speed regulation protection module reduces the forward speed of the mining machine, reduces tool wear, and reduces the risk of equipment failure in mixed coal seams.

[0015] Preferably, key features reflecting geological changes in coal seams are extracted from preprocessed image data, and the extracted features include the degree of texture distribution discreteness and edge features. Under the detection window, the extracted degree of texture distribution discreteness and edge features are analyzed to generate texture discrete values ​​and edge entropy values, respectively. The texture discrete value quantifies the unevenness and degree of discreteness of the texture distribution on the coal seam surface, reflecting the changes in the spatial distribution of different materials in the coal seam. The edge entropy value quantifies the complexity and irregularity of edge features in the coal seam image, reflecting the randomness of boundary distribution, density and morphology between different materials in the coal seam.

[0016] Preferably, under the detection window, the discrete degree of texture distribution is analyzed, and the specific steps of generating texture discrete values ​​are as follows:

[0017] In the detection window, the coal seam image is first divided into grids, and the local texture features are finely analyzed. Each grid represents the texture distribution of an area on the coal seam surface, and the local texture vector is calculated. The calculation expression is as follows:

[0018]

[0019] , where G θ (p,q) is the element in the gray-level co-occurrence matrix, T i,j is the local texture feature vector, that is, the local texture feature vector of the grid (i, j) area, i is the row index in the grid matrix, j is the column index in the grid matrix, and p and q are both grayscale values;

[0020] Based on the extracted local texture vector T i,j , calculate the texture feature gradient change between adjacent grids, that is, the difference in texture discreteness between different regions. The calculation expression is as follows:

[0021]

[0022] , where is the texture feature gradient, w x is the horizontal weight parameter, w y It is the vertical weight parameter, which indicates the weight coefficient of texture gradient change in the vertical direction;

[0023] After completing the calculation of the texture feature gradient between the grid regions, the texture discrete value is defined, and the calculation expression is as follows:

[0024]

[0025] , where TVDI is the texture discrete value, N is the total number of grids, is the texture feature gradient between the (i,j)th grid and the adjacent grids, is the maximum texture feature gradient.

[0026] Preferably, in the detection window, the edge features are analyzed and the specific steps of generating the edge entropy value are as follows:

[0027] In the detection window, edge detection is performed on the coal seam image to extract the edge information between different materials in the image and obtain the edge strength of the image. The calculation expression is as follows:

[0028]

[0029] , where is the gray value change rate of image I in the x direction, that is, the degree of change of pixel brightness along the x-axis. is the gray value change rate of image I in the y direction, that is, the degree of change of pixel brightness along the y axis, E(x,y) is the image edge intensity value, which indicates the edge intensity at the coordinate position (x,y) in the image;

[0030] Based on the extracted image edge intensity value E(x,y), the distribution of the edge area is analyzed by structural entropy to quantify the complexity and irregularity of the edge. The calculation expression is as follows:

[0031]

[0032] , where P k is the probability distribution of edge pixels, indicating the probability of occurrence of the kth edge pixel, E k is the edge strength value of the kth edge pixel, n is the total number of edge pixels, H edge is the edge structure entropy;

[0033] The edge structure entropy value H edge Normalization is performed to generate edge entropy values. The calculation expression is as follows:

[0034]

[0035] , where ECEV is the edge entropy value, D edge is the edge density, α is the amplification parameter, and β is the normalization weight parameter.

[0036] Preferably, after obtaining the texture discrete values ​​and edge entropy values ​​generated after analyzing the extracted key features, the texture discrete values ​​and edge entropy values ​​are input into a pre-learned machine learning model, and a coal seam change index is generated based on the machine learning model. The geological changes of the coal seam are intelligently evaluated through the coal seam change index.

[0037] Preferably, the coal seam change index generated when the geological change of the coal seam is intelligently evaluated by the pre-learned machine learning model is compared and analyzed with the pre-set coal seam change index reference threshold, and the coal seam geological change is divided. The specific division steps are as follows:

[0038] If the coal seam change index is greater than or equal to the set coal seam change index reference threshold, the current coal seam is classified as a mixed coal seam;

[0039] If the coal seam change index is less than the set coal seam change index reference threshold, the current coal seam is classified as a stable coal seam.

[0040] Preferably, for mixed coal seams, the specific steps of reducing the forward speed of the mining machine, reducing tool wear, and reducing the risk of equipment failure are as follows:

[0041] After the current coal seam is divided into a mixed coal seam, the speed adjustment coefficient is dynamically calculated according to the deviation between the coal seam change index CSVI and the reference threshold of the coal seam change index to reduce the forward speed of the mining machine. The calculation expression is as follows:

[0042]

[0043] , where CSVI ref is the reference threshold of the coal seam variation index, CSVI max is the maximum coal seam change index, k is the proportional adjustment factor of speed adjustment, γ is the speed adjustment coefficient, that is, the dynamic adjustment proportional coefficient of the forward speed of the mining machine,

[0044] Based on the preset forward speed and the speed adjustment coefficient γ obtained by dynamic calculation, the new forward speed of the mining machine is calculated in real time. The calculation expression is as follows:

[0045] V new =γ·V pre

[0046] , where V new is the adjusted forward speed of the mining machine, V pre is the preset forward speed of the mining machine;

[0047] The adjusted forward speed V new It is applied to mining machine operation, and continuously collects coal seam image data, updates the coal seam change index CSVI, and ensures that the mining machine speed is continuously adjusted with the changes in coal seam geology through real-time feedback and dynamic iterative optimization to achieve the best operating state. The calculation expression is as follows:

[0048] CSVI t+1 =f(CSVI t ,ΔV new )

[0049] , where CSVI t is the coal seam change index at the current moment, CSVI t+1 is the coal seam change index at the next moment, ΔV new is the adjustment amplitude of the forward speed, and the calculation expression is as follows:

[0050]

[0051] Through the feedback mechanism, the coal seam status is continuously updated, and the speed is dynamically adjusted according to real-time changes to form a closed-loop control.

[0052] In the above technical solution, the technical effects and advantages provided by the present invention are:

[0053] The present invention uses an AI video intelligent-assisted mine intelligent monitoring method to collect coal seam image data in real time and combine machine learning models to intelligently evaluate coal seam geological changes, thereby achieving dynamic classification of coal seams. For mixed coal seams, the system reduces the forward speed of the mining machine, reduces the friction and impact between the tool and the hard mineral, effectively reduces the wear rate and risk of tool collapse, and extends the service life of the tool. In addition, reducing the speed can also reduce the load of the mining machine in high-hardness areas, avoid equipment failures due to excessive tool wear, reduce downtime and maintenance time and costs, thereby ensuring the stable operation of the mining machine and improving equipment reliability and safety.

[0054] The present invention classifies coal seams into stable coal seams and mixed coal seams by intelligently evaluating coal seam changes, and dynamically adjusts the operating speed of the mining machine. In stable coal seams, the mining machine maintains the preset speed, efficiently completes the mining operation, and ensures that the production progress is not affected; while in mixed coal seams, the system automatically reduces the speed, effectively protects the cutter and equipment, and reduces failures and downtime. Through this intelligent dynamic regulation, the solution not only ensures efficient mining operations, but also achieves a balance between equipment protection and production efficiency, ultimately improving the overall economic benefits and stability of mining operations. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0056] Figure 1 The present invention is a method flow chart of the mine intelligent monitoring system based on AI video intelligent assistance. DETAILED DESCRIPTION

[0057] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as limited to the examples set forth herein; rather, these example embodiments are provided so that the description of the present disclosure will be more comprehensive and complete, and the concept of the example embodiments will be fully conveyed to those skilled in the art.

[0058] The present invention provides Figure 1 The AI ​​video intelligent assisted mine intelligent monitoring system shown in the figure includes an image data acquisition module, an image data preprocessing module, a geological feature extraction module, a machine learning intelligent evaluation module, a coal seam classification decision module, a stable coal seam operation module, and a mixed coal seam speed regulation protection module:

[0059] The image data acquisition module collects the coal seam image data in the direction of the mining machine in real time through the video monitoring system;

[0060] The coal seam image data in the direction of the mining machine is collected in real time through the video monitoring system. The specific steps include the following aspects: First, install high-resolution industrial cameras or sensors at key positions in the direction of the mining machine to ensure that the equipment can stably capture real-time image data of the coal seam; second, the camera continuously collects dynamic images of the coal seam at the working face through a preset viewing angle and frame rate (such as collecting several frames of images per second), covering the entire working area in front of the mining machine; then, the video data is transmitted to the central processing unit or data analysis server in real time through a wired or wireless transmission system (such as an industrial network or 5G communication) to ensure efficient data transmission and low latency; finally, the system records and preliminarily stores the received raw image data with a timestamp to form a continuous coal seam image stream, providing basic support for subsequent image preprocessing and geological feature analysis. This process requires ensuring that the video equipment has strong anti-interference capabilities and can operate stably in complex coal mine environments (such as insufficient light, dust, vibration, etc.), ensuring that the collected image data is clear, accurate, and without missing.

[0061] The image data preprocessing module preprocesses the acquired coal seam images to ensure that the acquired coal seam images are clear and accurately reflect the geological characteristics of the coal seams;

[0062] Image preprocessing steps include image denoising, lighting adjustment, contrast enhancement, edge detection and other operations. These processes can reduce noise interference, enhance valuable details in the image, and ensure that image quality can still be guaranteed in complex mining environments. For example, the operating environment of the mining machine may have insufficient lighting or there may be interference such as ore and dust in the image. Therefore, image denoising is required to ensure that the collected coal seam image is clear and can accurately reflect the geological characteristics of the coal seam. Image enhancement can also highlight the boundary between the coal seam and hard minerals, which is convenient for subsequent feature extraction.

[0063] The geological feature extraction module extracts key features reflecting the geological changes of coal seams from the preprocessed image data after preprocessing the acquired image data, and analyzes the extracted geological change features under the monitoring window to quantify the distribution and changes of geological features;

[0064] The key features reflecting the geological changes of the coal seam are extracted from the preprocessed image data. The extracted features include the discrete degree of texture distribution and edge features. Under the detection window, the extracted discrete degree of texture distribution and edge features are analyzed to generate texture discrete value and edge entropy value respectively. The texture discrete value quantifies the unevenness and discrete degree of texture distribution on the coal seam surface, reflecting the changes in the spatial distribution of different materials (such as coal, hard minerals, and rocks) in the coal seam. The edge entropy value quantifies the complexity and irregularity of edge features in the coal seam image, reflecting the randomness of boundary distribution, density and morphology between different materials (such as coal and hard minerals) in the coal seam.

[0065] The texture distribution shows obvious irregularity and discreteness, indicating that the current coal seam is in a mixed coal seam. This is mainly because there are significant differences in the physical properties and distribution characteristics of coal and hard minerals (such as quartz, feldspar, iron ore, etc.) in the mixed coal seam, resulting in the coal seam surface texture showing high inhomogeneity and irregularity. In a stable coal seam, the material of coal is relatively uniform, and the surface texture usually presents a smooth, continuous and regular distribution. However, in a mixed coal seam, the doping of hard minerals will form granular, blocky or irregular distribution areas, and the texture density, directionality and roughness of these areas are significantly different from the texture of coal. Therefore, in image analysis, the texture distribution will be discrete, that is, the texture characteristics of different areas show great differences and discontinuity. For example, the texture of the hard mineral area is rougher and the direction is random, while the texture of the coal seam area is relatively fine and uniform. This irregularity and discreteness of texture distribution actually reflects the inhomogeneity of the internal geological structure of the coal seam, indicating that hard minerals and coal are alternately distributed or intermixed in space, which is a typical feature of a mixed coal seam.

[0066] Under the detection window, the discrete degree of texture distribution is analyzed, and the specific steps to generate texture discrete values ​​are as follows:

[0067] In the detection window, the coal seam image is first divided into grids, and the local texture features are finely analyzed. Each grid represents the texture distribution of an area on the coal seam surface. Texture feature extraction uses a combination of grayscale co-occurrence matrix and directional features to describe the texture features in each grid, including directional changes, energy density and roughness distribution, and calculates the local texture vector. The calculation expression is as follows:

[0068]

[0069] , where G θ(p,q) is an element in the gray-level co-occurrence matrix, which indicates the co-occurrence relationship between the gray-level value p and the gray-level value q in the image at a certain direction angle θ. i,j is the local texture feature vector, that is, the local texture feature vector of the grid (i, j) area, i is the row index in the grid matrix, used to locate the grid in the vertical direction, j is the column index in the grid matrix, used to locate the grid in the horizontal direction, and is the comprehensive texture feature vector calculated for the grid in all directions and angles θ, p and q are both grayscale values;

[0070] Through the above steps, the directional texture information of each region can be extracted, and combined with the grayscale difference between pixels, the complexity and roughness of the local texture can be quantified.

[0071] Based on the extracted local texture vector T i,j , calculate the texture feature gradient change between adjacent grids, that is, the difference in texture discreteness between different regions. The calculation expression is as follows:

[0072]

[0073] , where is the texture feature gradient, w x is the horizontal direction (x direction) weight parameter, which represents the weight coefficient of the texture gradient change in the horizontal direction, w y It is the vertical direction (y direction) weight parameter, which represents the weight coefficient of texture gradient change in the vertical direction;

[0074] Through the above steps, the spatial "discreteness" and directional variation of the surface texture characteristics of the coal seam can be quantified. If the coal seam is a mixed coal seam, the interface between the hard minerals and the coal will show a large gradient change, resulting in The value increases.

[0075] After completing the calculation of the texture feature gradient between the grid regions, the texture discrete value is defined to comprehensively quantify the texture distribution unevenness and discreteness within the entire detection window. The entropy factor and gradient weight are introduced to further quantify and weight the texture difference. The calculation expression is as follows:

[0076]

[0077] , where TVDI is the texture discrete value, N is the total number of grids, is the texture feature gradient between the (i,j)th grid and the adjacent grids, is the maximum texture feature gradient.

[0078] When the TVDI value is large, it means that the surface texture of the coal seam changes dramatically and the distribution is highly discrete, reflecting that hard minerals, rocks and coal are mixed in space and the coal seam is in a mixed coal seam state. On the contrary, when the TVDI value is small, it means that the texture is uniform and changes smoothly, and the coal seam is relatively pure and stable.

[0079] Under the detection window, the larger the texture discrete value generated after analyzing the discrete degree of texture distribution, the greater the performance value, indicating that the current coal seam is in a mixed coal seam; conversely, the smaller the performance value, the current coal seam is in a stable coal seam. In a stable coal seam, the coal quality is relatively pure, the surface texture is evenly and continuously distributed, and the texture discrete value is small, showing a high degree of consistency. In a mixed coal seam, the doping of hard minerals such as quartz and feldspar will form granular and blocky high roughness areas on the image. The texture characteristics of these areas are in significant contrast with the smooth texture of coal, resulting in obvious discreteness and irregularity of the texture distribution in space, thereby significantly increasing the performance value of the texture discrete value. Therefore, the larger the value, the stronger the inhomogeneity of the coal seam, indicating that the content of mixed mineral components is high and the coal seam belongs to a mixed coal seam; conversely, the smaller the texture discrete value, the purer and more stable the coal seam.

[0080] The irregular and complex boundaries in the edge features indicate that the current coal seam is a mixed coal seam. This is mainly because the coal seam and hard minerals (such as quartz, feldspar, iron ore, etc.) have significant differences in physical properties, optical properties and structural morphology, which leads to a large number of high-contrast, irregular and dense boundaries in the image. Stable coal seams are usually composed of pure coal materials, with relatively uniform textures, few and regular edge features, and a relatively smooth structure as a whole. In mixed coal seams, due to the doping of hard minerals with completely different hardness, reflectivity and material, a large number of complex and discontinuous boundaries are formed at the junction of coal and hard minerals. These boundaries appear as areas with higher entropy values ​​in the image.

[0081] The irregularity of the boundaries reflects the uneven distribution of hard minerals in the coal seam and the diversity of particle sizes and shapes, indicating that the geological structure of the coal seam has changed significantly. In addition, the transition zone between hard minerals and coal is often more complex, with dense and irregular boundary lines, which further increases the complexity of edge features. Due to the interlaced distribution of hard minerals and coal in mixed coal seams, the number of boundaries increases and the shapes are diverse, such as granular, linear and even irregular fragments, which are typical characteristics of mixed coal seams.

[0082] Under the detection window, the specific steps for analyzing edge features and generating edge entropy values ​​are as follows:

[0083] In the detection window, edge detection is performed on the coal seam image to extract the edge information between different materials (such as coal and hard minerals) in the image and obtain the edge strength of the image. The calculation expression is as follows:

[0084]

[0085] , where It is the gray value change rate of image I in the x direction (horizontal direction), that is, the degree of change of pixel brightness along the x axis. is the gray value change rate of image I in the y direction (vertical direction), that is, the degree of change of pixel brightness along the y axis, E(x,y) is the image edge intensity value, which indicates the edge intensity at the coordinate position (x,y) in the image;

[0086] Through this process, the distribution and significance of the boundaries between different materials (such as hard minerals and coal) in the coal seam image can be obtained, providing a basis for the subsequent calculation of the edge entropy index.

[0087] Based on the extracted image edge strength value E(x,y), the distribution of the edge area is analyzed by structural entropy to quantify the complexity and irregularity of the edge. The calculation of edge structural entropy is based on the probability distribution of edge strength. The level of entropy reflects the randomness and uncertainty of boundary information. The calculation expression is as follows:

[0088]

[0089] , where P k It is the probability distribution of edge pixels, indicating the probability of occurrence of the kth edge pixel. It is obtained based on the statistical distribution of the edge intensity value E(x, y) of the image, reflecting the distribution frequency of the edge intensity value in different areas of the image. k is the edge intensity value of the kth edge pixel, indicating the significance of the edge pixel in the image, n is the total number of edge pixels, H edge is the edge structure entropy;

[0090] The edge structure entropy value H edge Normalization is performed to generate edge entropy values. The calculation expression is as follows:

[0091]

[0092] , where ECEV is the edge entropy value, D edge is the edge density, which indicates the ratio of edge pixels to total pixels in the unit detection window. α is the amplification parameter, which is used to emphasize the contribution of strong edges and suppress the influence of weak edges. β is the normalization weight parameter, which is used to adjust the weight of the influence of edge strength on the final edge entropy value.

[0093] Under the detection window, the larger the edge entropy value generated after analyzing the edge features, the more it indicates that the current coal seam is in a mixed coal seam. Conversely, the smaller the edge entropy value, the more it indicates that the coal seam is in a stable coal seam. This is because the edge entropy value quantifies the complexity and randomness of the edge features in the coal seam image, and it measures the uncertainty of the boundary information. In a stable coal seam, the coal quality is relatively uniform, the edges on the image are fewer and regularly distributed, and the entropy value of the boundary is low, indicating a simple structure and high consistency. In a mixed coal seam, due to the significant differences in reflectivity, texture, morphology, etc. between hard minerals and coal, a large number of irregular and complex edge features will be formed in the contact area between coal and hard minerals. These edges are dense, random and discontinuous in spatial distribution, resulting in increased uncertainty in edge information, thereby increasing the edge entropy value.

[0094] The machine learning intelligent evaluation module inputs the analyzed features into a pre-trained machine learning model and intelligently evaluates the geological changes of the coal seam through the feature quantification data;

[0095] After obtaining the texture discrete value and edge entropy value generated after analyzing the extracted key features, the texture discrete value and edge entropy value are input into the pre-learned machine learning model, and a coal seam change index is generated based on the machine learning model. The coal seam change index is used to intelligently evaluate the geological changes of the coal seam;

[0096] A pre-learned machine learning model refers to a model that is fully trained and obtained through a machine learning algorithm based on a large amount of historical data and training samples. This model can automatically learn the relationship between input features (such as texture discrete values ​​and edge entropy values) and output results (such as the geological change state of coal seams), and convert this law into a mathematical mapping to achieve intelligent evaluation of new data. In the assessment of coal seam geological changes, the model training process usually includes steps such as data collection, feature extraction, label calibration, model training and verification. By taking key features such as texture distribution unevenness and edge complexity in coal seam image data as input, and combining it with the labels of known coal seam states, the model learns the association between these features and coal seam changes, and finally forms an intelligent tool that can accurately identify and evaluate coal seam changes.

[0097] In specific implementations, pre-learned machine learning models may use a variety of algorithms, such as decision trees, support vector machines (SVM), random forests, or deep learning models (such as convolutional neural networks (CNNs)). These algorithms continuously optimize model parameters through training with large amounts of data to ensure their understanding and generalization of coal seam change characteristics. In practical applications, the model does not need to be retrained. It only needs to input new texture discrete values ​​and edge entropy values ​​to quickly generate a coal seam change index based on the knowledge learned during training, thereby evaluating the geological changes of the coal seams.

[0098] The pre-learned machine learning model plays a core role in coal seam geological assessment. Its main function is to realize the intelligent quantification and accurate assessment of complex coal seam geological characteristics. By taking the extracted texture discrete values ​​and edge entropy values ​​as input, the machine learning model can comprehensively analyze the contribution of these features to coal seam geological changes and generate a coal seam change index (CSVI). This index quantifies the stability and degree of change of coal seams through mathematical mapping, thereby reflecting the distribution differences of different materials such as coal and hard minerals and rocks in coal seams. The introduction of preset proportional coefficients y1y_1 and y2y_2 further ensures that the model reasonably distributes the weights of different features during the evaluation process, and can dynamically adjust the importance of features according to actual geological conditions.

[0099] In addition, the machine learning model has been verified and optimized through a large amount of calibration data during the training process, and has strong generalization capabilities and can adapt to geological changes in different coal mine environments. For example, in complex coal seams, the model can capture subtle changes in texture and edge features, quickly determine whether the current coal seam is a stable coal seam or a mixed coal seam, and provide a scientific basis for the mining machine to dynamically adjust the operating parameters. Compared with traditional manual judgment or simple algorithm analysis, the pre-learned model has the advantages of high precision, high real-time performance and automation, which greatly improves the level of intelligence in coal mining and reduces the risk of equipment wear and operating costs.

[0100] The machine learning model is not limited here, and any machine learning model that can generate the coal seam variation index CSVI after comprehensive analysis of the texture discrete value TVDI and the edge entropy value ECEV is acceptable. To realize the technical solution of the present invention, the present invention provides a specific implementation method;

[0101] The formula for generating the coal seam variation index CSVI is as follows:

[0102]

[0103] , where y 1 ,y 2 are the preset proportional coefficients of the texture discrete value TVDI and the edge entropy value ECEV, and y 1 ,y 2 Both are greater than 0.

[0104] The preset proportionality factor y here 1 and 2 It is used to adjust the weight or contribution of texture discrete value TVDI and edge entropy value ECEV in the generation process of coal seam variation index CSVI. Specifically, these two coefficients represent the relative importance of these two features in the model evaluation process. For example, if the unevenness of texture distribution (texture discrete value TVDI) in a coal seam environment has a more significant impact on coal seam geological changes, then y 1The value of y will be relatively large, indicating that the texture discrete value TVDI plays a dominant role in the calculation of the coal seam variation index CSVI; on the contrary, if the edge complexity (edge ​​entropy value ECEV) can better reflect the distribution of hard minerals in the coal seam, then y 2 The value of will be higher, giving greater weight to the edge entropy value ECEV. These preset proportionality coefficients y 1 and 2 The determination of is usually optimized during the training process of the machine learning model, based on a large number of historical data samples and label learning. The model analyzes the relationship between the texture discrete value and the edge entropy value and the actual state of the coal seam change, and continuously iterates and adjusts these two coefficients to achieve the best weight distribution, thereby ensuring that the final generated coal seam change index CSVI can accurately reflect the geological changes of the coal seam.

[0105] It can be seen from the coal seam change index that, under the detection window, the larger the texture discrete value expression value generated after analyzing the discrete degree of texture distribution, and the larger the edge entropy value expression value generated after analyzing the edge features, the larger the coal seam change index expression value generated when the geological changes of the coal seam are intelligently evaluated through the pre-learned machine learning model, indicating that the current coal seam is in a mixed coal seam, otherwise, it indicates that the current coal seam is in a stable coal seam.

[0106] The coal seam classification decision module classifies coal seam geological changes into stable coal seams and mixed coal seams based on the evaluation results of the machine learning model;

[0107] The coal seam change index generated by the pre-learned machine learning model when the coal seam geological change is intelligently evaluated is compared with the pre-set coal seam change index reference threshold to divide the coal seam geological change. The specific division steps are as follows:

[0108] If the coal seam change index is greater than or equal to the set coal seam change index reference threshold, the current coal seam is classified as a mixed coal seam;

[0109] Mixed coal seams are areas of the seam where there are more hard minerals or large differences in mineral hardness, which may lead to accelerated tool wear or increased risk of equipment failure.

[0110] If the coal seam change index is less than the set coal seam change index reference threshold, the current coal seam is classified as a stable coal seam.

[0111] A stable coal seam is an area in the seam where the coal is pure, the hard mineral content is low or evenly distributed, and the mining machine can continue to operate at a preset speed.

[0112] Stable coal seam operation module: For stable coal seams, the ore layer is uniform and the coal quality is relatively pure. The mining machine mines at the preset forward speed to maintain mining efficiency;

[0113] For stable coal seams, since the ore seams are uniform and the coal quality is relatively pure, the main purpose of the mining machine mining at the preset forward speed is to maximize operating efficiency and optimize resource utilization. The coal quality in stable coal seams is pure and the hardness is consistent. The cutting resistance between the cutter and the coal seam is small, and the equipment wear is relatively low. Therefore, there is no need to adjust the speed, and it can run stably at a fixed and efficient speed. This operation method can maintain the continuity and high productivity of the mining machine, reduce unnecessary downtime and loss of operating time, and thus improve the efficiency of coal collection. In addition, the preset forward speed is based on a comprehensive consideration of equipment performance and coal seam characteristics, ensuring that energy consumption is minimized and equipment loss is optimized under stable conditions, so as to achieve the economy and sustainability of coal mining. Therefore, for stable coal seams, maintaining a fixed speed operation can maximize mining efficiency and ensure the efficient achievement of production goals while ensuring equipment safety.

[0114] Mixed coal seam speed regulation protection module, for mixed coal seams, reduces the forward speed of the mining machine, reduces tool wear, and reduces the risk of equipment failure;

[0115] For mixed coal seams, the specific steps to reduce the forward speed of the mining machine, reduce tool wear, and reduce the risk of equipment failure are as follows:

[0116] After the current coal seam is divided into a mixed coal seam, the speed adjustment coefficient is dynamically calculated according to the deviation between the coal seam change index CSVI and the reference threshold of the coal seam change index to reduce the forward speed of the mining machine. The calculation expression is as follows:

[0117]

[0118] , where CSVI ref is the reference threshold of the coal seam variation index, CSVI max is the maximum coal seam variation index, representing the complexity of the extreme mixed coal seam, k is the proportional adjustment factor of speed adjustment, controlling the adjustment amplitude of the forward speed of the mining machine, γ is the speed adjustment coefficient, i.e., the dynamic adjustment proportional coefficient of the forward speed of the mining machine,

[0119] Based on the preset forward speed and the speed adjustment coefficient γ obtained by dynamic calculation, the new forward speed of the mining machine is calculated in real time to reduce the impact force between the tool and the hard mineral, reduce the wear rate and the risk of equipment failure. The calculation expression is as follows:

[0120] V new =γ·V pre

[0121] , where V new is the adjusted forward speed of the mining machine, V pre is the preset forward speed of the mining machine;

[0122] The adjusted forward speed V new It is applied to mining machine operation, and continuously collects coal seam image data, updates the coal seam change index CSVI, and ensures that the mining machine speed is continuously adjusted with the changes in coal seam geology through real-time feedback and dynamic iterative optimization to achieve the best operating state. The calculation expression is as follows:

[0123] CSVI t+1 =f(CSVI t ,ΔV new )

[0124] , where CSVI t is the coal seam change index at the current moment, CSVI t+1 is the coal seam change index at the next moment, ΔV new It is the adjustment amplitude of the forward speed, which means the adjustment amount of the forward speed of the mining machine calculated based on the complexity of the coal seam at time t, ensuring that the speed of the mining machine adapts to the dynamic changes of the coal seam conditions. The calculation expression is as follows:

[0125]

[0126] Through the feedback mechanism, the coal seam status is continuously updated and the speed is dynamically adjusted according to real-time changes to form a closed-loop control. If the coal seam status returns to stability and the coal seam change index CSVI is lower than the threshold, the preset forward speed of the mining machine will be gradually restored to ensure maximum efficiency.

[0127] For mixed coal seams, the main purpose of reducing the forward speed of the mining machine is to protect the equipment tools, reduce the risk of equipment failure, and optimize the safety and economy of the coal mining process. Mixed coal seams are doped with minerals that are significantly harder than coal (such as quartz, feldspar, iron ore, etc.). These hard minerals will increase the wear and impact of the tools during the mining process, resulting in a shortened tool life and even causing cracking or breakage. If the mining machine continues to run at the preset high speed, the intense friction and impact between the tools and the hard minerals will increase significantly, which will not only lead to faster tool wear, but may also cause additional loads on the transmission system, hydraulic system and power system of the mining machine, increasing the risk of equipment failure. Frequent equipment failures will lead to downtime for maintenance, which will not only affect the mining progress, but also increase maintenance costs and production losses.

[0128] By reducing the forward speed of the mining machine, the contact impact and friction intensity between the tool and the hard mineral can be effectively reduced, the wear rate of the tool can be slowed down, and the service life of the tool can be extended. In addition, the deceleration operation reduces the load pressure on the equipment, which helps to reduce problems such as motor overload and excessive pressure in the hydraulic system, reduce the overall failure probability of the equipment system, and ensure the stable operation of the mining machine. At the same time, although appropriately reducing the speed may affect the operating efficiency in the short term, in the long run, it can significantly reduce the unplanned downtime caused by equipment damage or shutdown maintenance, and ensure the continuity and economy of mining operations. In addition, deceleration operation can also improve the safety of mining operations and reduce the safety risks of the working surface caused by equipment failure, especially in high-risk areas under complex geological conditions.

[0129] In summary, reducing the forward speed of the mining machine is a dynamic protection measure that can achieve a balance between equipment protection and operating efficiency under mixed coal seam conditions, minimize tool wear, reduce the risk of equipment failure, and ensure the stability and economic benefits of the coal mining process.

[0130] The present invention uses an AI video intelligent-assisted mine intelligent monitoring method to collect coal seam image data in real time and combine machine learning models to intelligently evaluate coal seam geological changes, thereby achieving dynamic classification of coal seams. For mixed coal seams, the system reduces the forward speed of the mining machine, reduces the friction and impact between the tool and the hard mineral, effectively reduces the wear rate and risk of tool collapse, and extends the service life of the tool. In addition, reducing the speed can also reduce the load of the mining machine in high-hardness areas, avoid equipment failures due to excessive tool wear, reduce downtime and maintenance time and costs, thereby ensuring the stable operation of the mining machine and improving equipment reliability and safety.

[0131] The present invention classifies coal seams into stable coal seams and mixed coal seams by intelligently evaluating coal seam changes, and dynamically adjusts the operating speed of the mining machine. In stable coal seams, the mining machine maintains the preset speed, efficiently completes the mining operation, and ensures that the production progress is not affected; while in mixed coal seams, the system automatically reduces the speed, effectively protects the cutter and equipment, and reduces failures and downtime. Through this intelligent dynamic regulation, the solution not only ensures efficient mining operations, but also achieves a balance between equipment protection and production efficiency, ultimately improving the overall economic benefits and stability of mining operations.

[0132] The above description is only by way of illustration of certain exemplary embodiments of the present invention. It is undoubted that those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. The mine intelligent monitoring system based on AI video intelligent assistance is characterized by: It includes image data acquisition module, image data preprocessing module, geological feature extraction module, machine learning intelligent evaluation module, coal seam classification decision module, stable coal seam operation module and mixed coal seam speed regulation protection module: The image data acquisition module collects the coal seam image data in the direction of the mining machine in real time through the video monitoring system; The image data preprocessing module preprocesses the acquired coal seam images to ensure that the acquired coal seam images are clear and accurately reflect the geological characteristics of the coal seams; The geological feature extraction module extracts key features reflecting the geological changes of coal seams from the preprocessed image data after preprocessing the acquired image data, and analyzes the extracted geological change features under the monitoring window to quantify the distribution and changes of geological features; The machine learning intelligent evaluation module inputs the analyzed features into a pre-trained machine learning model and intelligently evaluates the geological changes of the coal seam through the feature quantification data; The coal seam classification decision module classifies coal seam geological changes into stable coal seams and mixed coal seams based on the evaluation results of the machine learning model; Stable coal seam operation module: For stable coal seams, the ore layer is uniform and the coal quality is relatively pure. The mining machine mines at the preset forward speed to maintain mining efficiency; The mixed coal seam speed regulation protection module reduces the forward speed of the mining machine, reduces tool wear and reduces the risk of equipment failure in mixed coal seams.

2. The AI ​​video intelligent assisted mine intelligent monitoring system according to claim 1 is characterized in that: The key features reflecting the geological changes of coal seams are extracted from the preprocessed image data. The extracted features include the discrete degree of texture distribution and edge features. Under the detection window, the extracted discrete degree of texture distribution and edge features are analyzed to generate texture discrete value and edge entropy value respectively. The texture discrete value quantifies the unevenness and discrete degree of texture distribution on the coal seam surface, reflecting the changes in the spatial distribution of different materials in the coal seam. The edge entropy value quantifies the complexity and irregularity of edge features in the coal seam image, reflecting the randomness of boundary distribution, density and morphology between different materials in the coal seam.

3. The AI ​​video intelligent assisted mine intelligent monitoring system according to claim 2 is characterized in that: Under the detection window, the discrete degree of texture distribution is analyzed, and the specific steps to generate texture discrete values ​​are as follows: In the detection window, the coal seam image is first divided into grids, and the local texture features are finely analyzed. Each grid represents the texture distribution of an area on the coal seam surface, and the local texture vector is calculated. The calculation expression is as follows: ， In the formula, G θ (p,q) is the element in the gray-level co-occurrence matrix, T i,j is the local texture feature vector, that is, the local texture feature vector of the grid (i, j) area, i is the row index in the grid matrix, j is the column index in the grid matrix, and p and q are both grayscale values; Based on the extracted local texture vector T i,j , calculate the texture feature gradient change between adjacent grids, that is, the difference in texture discreteness between different regions. The calculation expression is as follows: ， In the formula, is the texture feature gradient, w x is the horizontal weight parameter, w y It is the vertical weight parameter, which indicates the weight coefficient of texture gradient change in the vertical direction; After completing the calculation of the texture feature gradient between the grid regions, the texture discrete value is defined, and the calculation expression is as follows: ， Where TVDI is the texture discrete value, N is the total number of grids, is the texture feature gradient between the (i,j)th grid and the adjacent grids, is the maximum texture feature gradient.

4. The AI ​​video intelligent assisted mine intelligent monitoring system according to claim 1 is characterized in that: Under the detection window, the specific steps for analyzing edge features and generating edge entropy values ​​are as follows: In the detection window, edge detection is performed on the coal seam image to extract the edge information between different materials in the image and obtain the edge strength of the image. The calculation expression is as follows: ， In the formula, is the gray value change rate of image I in the x direction, that is, the degree of change of pixel brightness along the x-axis. is the gray value change rate of image I in the y direction, that is, the degree of change of pixel brightness along the y axis, E(x,y) is the image edge intensity value, which indicates the edge intensity at the coordinate position (x,y) in the image; Based on the extracted image edge intensity value E(x,y), the distribution of the edge area is analyzed by structural entropy to quantify the complexity and irregularity of the edge. The calculation expression is as follows: ， Where P k is the probability distribution of edge pixels, indicating the probability of occurrence of the kth edge pixel, E k is the edge strength value of the kth edge pixel, n is the total number of edge pixels, H edge is the edge structure entropy; The edge structure entropy value H edge Normalization is performed to generate edge entropy values. The calculation expression is as follows: ， Where ECEV is the edge entropy value, D edge is the edge density, α is the amplification parameter, and β is the normalization weight parameter.

5. The AI ​​video intelligent assisted mine intelligent monitoring system according to claim 2 is characterized in that: After obtaining the texture discrete value and edge entropy value generated after analyzing the extracted key features, the texture discrete value and edge entropy value are input into the pre-learned machine learning model, and a coal seam change index is generated based on the machine learning model. The coal seam change index is used to intelligently evaluate the geological changes of the coal seam.

6. The AI ​​video intelligent assisted mine intelligent monitoring system according to claim 5 is characterized in that: The coal seam change index generated by the pre-learned machine learning model when the coal seam geological change is intelligently evaluated is compared with the pre-set coal seam change index reference threshold to divide the coal seam geological change. The specific division steps are as follows: If the coal seam change index is greater than or equal to the set coal seam change index reference threshold, the current coal seam is classified as a mixed coal seam; If the coal seam change index is less than the set coal seam change index reference threshold, the current coal seam is classified as a stable coal seam.

7. The AI ​​video intelligent assisted mine intelligent monitoring system according to claim 6 is characterized in that: For mixed coal seams, the specific steps to reduce the forward speed of the mining machine, reduce tool wear, and reduce the risk of equipment failure are as follows: After the current coal seam is divided into a mixed coal seam, the speed adjustment coefficient is dynamically calculated according to the deviation between the coal seam change index CSVI and the reference threshold of the coal seam change index to reduce the forward speed of the mining machine. The calculation expression is as follows: ， Where, CSVI ref is the reference threshold of the coal seam variation index, CSVI max is the maximum coal seam change index, k is the proportional adjustment factor of speed adjustment, γ is the speed adjustment coefficient, that is, the dynamic adjustment proportional coefficient of the forward speed of the mining machine, Based on the preset forward speed and the speed adjustment coefficient γ obtained by dynamic calculation, the new forward speed of the mining machine is calculated in real time. The calculation expression is as follows: V new =γ·V pre , Where V new is the adjusted forward speed of the mining machine, V pre is the preset forward speed of the mining machine; The adjusted forward speed V new It is applied to mining machine operation, and continuously collects coal seam image data, updates the coal seam change index CSVI, and ensures that the mining machine speed is continuously adjusted with the changes in coal seam geology through real-time feedback and dynamic iterative optimization to achieve the best operating state. The calculation expression is as follows: CSVI t+1 =f(CSVI t ,ΔV new ), Where, CSVI t is the coal seam change index at the current moment, CSVI t+1 is the coal seam change index at the next moment, ΔV new is the adjustment amplitude of the forward speed, and the calculation expression is as follows: Through the feedback mechanism, the coal seam status is continuously updated, and the speed is dynamically adjusted according to real-time changes to form a closed-loop control.

Images