Dynamic visual observation method based on transparent coal rock material and internal cracks of transparent coal rock material

Abstract

The invention discloses a dynamic visual observation method based on a transparent coal rock material and internal cracks thereof, and the method comprises the steps: constructing a transparent coal rock three-dimensional space coordinate system, carrying out the meshing, carrying out the multi-scale feature extraction through employing an optimization SSD algorithm, simulating a light propagation path according to optical parameters, carrying out the modeling of a crack initiation and propagation time sequence through combining with mechanical parameters, and carrying out the visual observation of the internal cracks of the transparent coal rock material. Accurate crack positioning and contour extraction are achieved through an algorithm, and finally crack parameters and the three-dimensional model are fused to present an observation result. Meanwhile, the optimized SSD algorithm adopts mechanisms such as a feature fusion model and a self-adaptive adjustment convolution kernel, special models are constructed in links such as light propagation and crack evolution, and transparent coal rock multi-parameter calculation is combined. According to the method, high-precision dynamic visual observation of the internal cracks of the transparent coal rock is realized.

Description

Technical Field

[0001] The present invention relates to the field of transparent coal rock visualization, and in particular to a method for dynamic visualization observation of transparent coal rock materials and internal cracks thereof. Background Art

[0002] In fields such as coal mining and geomechanics, a deep understanding of the initiation and propagation of cracks within coal rock is crucial for ensuring mining safety and improving resource utilization. Transparent coal rock materials, which can simulate the physical and optical properties of real coal rock, are ideal for studying its internal structure. However, traditional observation methods are insufficient for accurately studying the dynamic changes in cracks, and more advanced observation methods are urgently needed.

[0003] Existing technologies have significant deficiencies. On the one hand, the observation methods lack the comprehensive application of transparent coal rock material properties. Most traditional methods rely only on a single parameter or a simple model, and do not fully combine the elastic modulus, Poisson's ratio, transmittance and other multi-dimensional physical and optical parameters of transparent coal rock. They cannot accurately describe the propagation law of light in cracked transparent coal rock, resulting in deviations in the judgment of crack location and morphology. On the other hand, dynamic observation capabilities are limited. Traditional methods often use static or low-frequency observations, which make it difficult to capture the rapid expansion process of cracks in complex stress environments. There are also technical bottlenecks in multi-scale feature extraction and three-dimensional spatial modeling. It is impossible to present the dynamic evolution of cracks in real time and continuously, which greatly limits the in-depth study of the coal rock fracture mechanism. Summary of the Invention

[0004] In order to overcome the shortcomings and deficiencies of the prior art, the present invention provides a method for dynamic visualization observation of transparent coal rock materials and internal cracks thereof.

[0005] The technical solution adopted by the present invention is based on a method for dynamic visualization observation of transparent coal rock materials and internal cracks thereof, which includes:

[0006] Step S1: constructing a three-dimensional spatial coordinate system for the transparent coal rock material, performing spatial gridding processing on the transparent coal rock sample based on the physical parameters of the transparent coal rock, such as the elastic modulus, Poisson's ratio, and density, and dividing the transparent coal rock sample into discrete units with three-dimensional spatial coordinates;

[0007] Step S2: Using the optimized SSD algorithm, a multi-scale feature extraction mechanism is established for the surface features and internal structure of discrete units of transparent coal rock. Through feature extraction and fusion of convolutional neural network layers, a feature map containing information about potential crack areas is obtained.

[0008] Step S3: Based on the optical parameters of the transparent coal rock, such as transmittance and refractive index, a light propagation simulation is performed on the characteristic graph in a three-dimensional spatial coordinate system to construct the interaction between the light and the discrete units of the transparent coal rock, and to determine the refraction, reflection, and scattering paths caused by cracks when the light propagates inside the transparent coal rock;

[0009] Step S4: Based on the light propagation simulation results, combined with the crack growth rate and crack tip stress intensity factor mechanical parameters of the transparent coal rock, the crack initiation and growth process are modeled in a three-dimensional dynamic visualization observation model to generate a three-dimensional model sequence containing crack morphologies at different time nodes;

[0010] Step S5: Using the non-maximum suppression and bounding box regression mechanism in the optimized SSD algorithm, the crack area in the 3D model sequence is accurately located and contour extracted to obtain the geometric size and spatial position parameters of the crack;

[0011] Step S6: The geometric dimensions, spatial position parameters and three-dimensional models of crack morphology at different time nodes are integrated and presented in a three-dimensional dynamic visualization observation model to form a dynamic visualization observation result of cracks inside transparent coal rock.

[0012] Furthermore, in step S2, the optimized SSD algorithm adopts an improved feature pyramid structure to construct the following feature fusion model:

[0013]

[0014] Among them, F merged is the fused feature map; are feature maps of different levels respectively; α, β, and γ are fusion weights, whose values ​​are determined by the adaptive learning algorithm based on the elastic modulus and Poisson's ratio physical parameters of transparent coal rock, and are used to balance the contribution of feature maps of different levels to crack feature extraction.

[0015] Furthermore, in step S3, the propagation path of light inside the transparent coal rock is determined using the following model: in, is the position vector of the light in the nth discrete unit body; is the position vector of the light in the n+1th discrete unit body; Δs is the distance that the light propagates between adjacent discrete units, which is determined according to the density and transmittance parameters of the transparent coal rock; is the propagation direction vector of the light in the nth discrete unit body, which is determined by the refractive index of the transparent coal rock and the geometric shape of the crack surface.

[0016] Furthermore, in step S4, the time series modeling of the crack propagation process adopts the following dynamic evolution model:

[0017]

[0018] Where C(t) is the crack profile at time t; C(t+Δt) is the crack profile at time t+Δt; δ is the crack growth increment coefficient, which is determined according to the crack growth rate of transparent coal rock; K I (t) is the stress intensity factor at the crack tip at time t; K IC is the fracture toughness of transparent coal rock; is the normal vector of the crack tip at time t.

[0019] Furthermore, in step S5, the non-maximum suppression mechanism used for crack area positioning introduces the grayscale distribution parameters of the transparent coal rock for optimization, and the optimized non-maximum suppression threshold determination model is as follows: T = μ + λ·σ, where T is the non-maximum suppression threshold; μ is the mean of the grayscale value of the transparent coal rock; σ is the standard deviation of the grayscale value of the transparent coal rock; λ is the adjustment coefficient, which is determined according to the transmittance of the transparent coal rock and the contrast between the crack and the background.

[0020] Furthermore, in step S1, the spatial gridding process of the transparent coal rock sample is combined with the particle size distribution parameters of the transparent coal rock, and a variable resolution gridding model is adopted, specifically as follows: N cell =f(d particle ,ρ density ), where N cell is the number of units after meshing; d particle is the average particle size of transparent coal rock particles; ρ density is the density of transparent coal rock; f is the mapping function established according to the physical properties of transparent coal rock, which is used to determine the grid resolution of different areas.

[0021] Furthermore, in step S2, in the convolutional neural network layer of the optimized SSD algorithm, the convolution kernel parameters are adjusted according to the texture feature parameters of the transparent coal rock, using the following adaptive adjustment model: Among them, w ij is the adjusted convolution kernel parameter; is the initial convolution kernel parameter; Δw is the parameter adjustment step; g(θ texture ) is the texture direction of transparent coal rock θ texture Related functions are used to enable the convolution kernel to better extract crack texture features.

[0022] Furthermore, in step S4, the rendering of the crack morphology in the three-dimensional dynamic visualization observation model adopts an illumination model based on the optical parameters of transparent coal rock, specifically: Where, I is the final illumination intensity on the crack surface; k a 、k d 、k sare the ambient light reflection coefficient, diffuse reflection coefficient, and specular reflection coefficient, respectively, which are determined according to the transmittance and refractive index of transparent coal rock; I a , I d , I s They are ambient light intensity, diffuse light intensity, and specular light intensity respectively; is the crack surface normal vector; is the light direction vector; is the viewpoint direction vector; is the direction vector of the reflected light; n is the glossiness index, which is determined according to the surface roughness of the transparent coal rock.

[0023] Furthermore, in step S6, the fusion presentation of the three-dimensional dynamic visualization observation results adopts a multimodal data fusion model, which combines the geometric parameters and mechanical parameters of the transparent coal rock as follows: final =ω1·V geometry +ω2·V mechanics , where V final is the final three-dimensional dynamic visualization observation result; V geometry V is a visualization model based on the crack geometry and spatial position; mechanics is a visualization model based on the mechanical properties of cracks; ω1 and ω2 are fusion weights, which are determined by the weighted average algorithm according to the elastic modulus and crack growth rate parameters of transparent coal rock.

[0024] Beneficial effects: The present invention proposes a dynamic visualization observation method based on transparent coal rock materials and their internal cracks. The method spatially grids the sample based on physical parameters such as elastic modulus and Poisson's ratio, simulates light propagation according to optical parameters such as transmittance and refractive index, constructs the interaction between light and discrete units, accurately determines the refraction, reflection and scattering paths of light caused by cracks, and achieves accurate judgment of the crack position and morphology. In response to the problem of limited dynamic observation capabilities, this method combines mechanical parameters such as crack propagation rate and stress intensity factor to perform time-series modeling of crack initiation and propagation processes, generates a three-dimensional model sequence at different time nodes, and uses the optimized SSD algorithm to achieve precise positioning and contour extraction of the crack area, thereby presenting the full picture of the dynamic evolution of the crack in real time and continuously. In addition, the optimized SSD algorithm improves the multi-scale feature extraction capability by constructing a feature fusion model, adaptively adjusting the convolution kernel parameters, etc.; the three-dimensional dynamic visualization observation model enhances the comprehensiveness of visualization effects and information presentation with the help of illumination models, multimodal data fusion models, etc. This method systematically solves existing technical difficulties and provides an efficient and accurate technical means for the study of internal cracks in transparent coal rocks, which will help promote technological development and theoretical research in fields such as coal mining and geomechanics. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 is a flow chart of the method steps of the present invention;

[0026] Figure 2 This is a diagram of the unit composition of the method implementation of the present invention. DETAILED DESCRIPTION

[0027] It should be noted that, unless there is a conflict, the embodiments in this application and the features described in the embodiments can be combined with each other. The application is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0028] like Figure 1 As shown, based on the transparent coal rock material and the dynamic visualization observation method of the internal cracks thereof, the method includes:

[0029] Step S1: constructing a three-dimensional spatial coordinate system for the transparent coal rock material, performing spatial gridding processing on the transparent coal rock sample based on the physical parameters of the transparent coal rock, such as the elastic modulus, Poisson's ratio, and density, and dividing the transparent coal rock sample into discrete units with three-dimensional spatial coordinates;

[0030] Specifically, at the initial stage of observing cracks inside transparent coal rock, step S1 focuses on building a basic spatial framework. First, a three-dimensional spatial coordinate system for the transparent coal rock material is constructed. This coordinate system provides a spatial reference for all subsequent observations and analyses. Based on the physical parameters of the transparent coal rock, such as the elastic modulus, Poisson's ratio, and density, the transparent coal rock sample is spatially gridded. The elastic modulus reflects the material's ability to resist elastic deformation, the Poisson's ratio reflects the relationship between transverse strain and longitudinal strain, and the density is related to the material's mass distribution characteristics. Based on these parameters, the transparent coal rock sample is divided into discrete units with three-dimensional spatial coordinates. This discretization method enables the study of complex coal rock samples to be refined to each unit, laying the foundation for the subsequent precise analysis of the state of cracks in different locations and different mechanical environments, and can more accurately capture the subtle changes in cracks in various regions within the coal rock.

[0031] In terms of implementation, when constructing a three-dimensional spatial coordinate system, a fixed origin and three mutually perpendicular coordinate axes are determined as a reference for spatial positioning. For physical parameters, accurate elastic modulus, Poisson's ratio and density data of transparent coal rock are obtained through experimental measurements and other methods. Then, based on these data, a specific grid division algorithm is used to divide the sample into many discrete units. Each unit is assigned a clear three-dimensional spatial coordinate to ensure that its position can be accurately identified in subsequent research. In this way, the originally continuous transparent coal rock sample is converted into a discrete, analyzable set of units, providing a structured data foundation for in-depth research on the initiation and expansion of cracks in the coal rock, making the study of the complex internal structure of coal rock more organized and operational.

[0032] Step S2: Using the optimized SSD algorithm, a multi-scale feature extraction mechanism is established for the surface features and internal structure of discrete units of transparent coal rock. Through feature extraction and fusion of convolutional neural network layers, a feature map containing information about potential crack areas is obtained.

[0033] Specifically, step S2 mainly utilizes the optimized SSD algorithm to achieve efficient extraction of features of discrete units of transparent coal rock. The surface features and internal structures of discrete units of transparent coal rock are complex and diverse, and the presence of cracks will change the manifestation of these features. The optimized SSD algorithm establishes a multi-scale feature extraction mechanism, and obtains a feature map containing information about potential crack areas through feature extraction and fusion of convolutional neural network layers. In this process, the convolutional neural network layer processes the image data of the unit body through convolution kernels of different sizes and parameters, capturing its features at different scales. Smaller convolution kernels can extract local detail features, while larger convolution kernels can obtain more macroscopic structural features. Through layer-by-layer processing of multi-layer convolutional neural networks, features of different scales are fused to comprehensively capture the feature information of areas where cracks may exist, providing a key basis for the subsequent accurate identification of crack location and morphology.

[0034] In terms of implementation, the transparent coal rock discrete unit image data divided by step S1 is first input into the optimized SSD algorithm. The algorithm starts the convolutional neural network layer and performs convolution operations on the image data layer by layer according to the preset convolution kernel parameters and network structure. Each layer of convolution operation will extract features of a specific scale. As the number of network layers increases, the extracted features gradually transition from simple underlying features such as edges and textures to more complex semantic features. In the feature fusion stage, the algorithm combines these features according to the importance and relevance of features at different levels to generate a feature map containing information about potential crack areas. In this way, key features related to cracks can be efficiently screened out from complex coal rock image data, greatly improving the accuracy and efficiency of crack detection and reducing misjudgments and missed judgments due to the complexity of the coal rock material itself.

[0035] Step S3: Based on the optical parameters of the transparent coal rock, such as transmittance and refractive index, a light propagation simulation is performed on the characteristic graph in a three-dimensional spatial coordinate system to construct the interaction between the light and the discrete units of the transparent coal rock, and to determine the refraction, reflection, and scattering paths caused by cracks when the light propagates inside the transparent coal rock;

[0036] Specifically, step S3 simulates the propagation of light on the characteristic map in the constructed three-dimensional space coordinate system based on the optical parameters of the transparent coal rock, such as the transmittance and refractive index. The transmittance determines the degree to which light can penetrate the transparent coal rock, while the refractive index affects the change in the propagation direction of light inside the coal rock. When cracks exist inside the coal rock, the light will be refracted, reflected, and scattered due to the existence of the crack interface during propagation. By constructing the interaction between light and discrete units of transparent coal rock and simulating the propagation path of light inside the coal rock, the various optical changes caused by cracks can be accurately determined. This helps to analyze the existence and morphology of cracks from an optical perspective, and uses the changes in light propagation characteristics to present the position and geometric shape of cracks inside the coal rock in the form of optical phenomena, providing an intuitive optical basis for the subsequent visualization and analysis of cracks.

[0037] During the implementation process, the scope of the area where cracks may exist is first determined based on the characteristic map obtained in step S2. Then, a light propagation model is established based on the pre-measured optical parameters of the transparent coal rock, such as the transmittance and refractive index. In the three-dimensional spatial coordinate system, light is emitted to the transparent coal rock sample from different angles and positions, and the refraction, reflection, and scattering of the light when it encounters discrete units, especially when passing through the crack area, are calculated based on optical principles. By calculating the propagation path and optical changes of the light inside the coal rock point by point, a complete light propagation simulation result is constructed. These results record in detail the propagation trajectory of light inside the coal rock and the optical effects caused by cracks, so that the influence of the originally invisible cracks on the light can be quantified and visualized, providing important data support for in-depth analysis of crack characteristics.

[0038] Step S4: Based on the light propagation simulation results and combined with mechanical parameters such as the crack growth rate and crack tip stress intensity factor of the transparent coal rock, a time series modeling of the crack initiation and growth process is performed in a three-dimensional dynamic visualization observation model to generate a three-dimensional model sequence containing crack morphologies at different time nodes;

[0039] Specifically, step S4 is based on the light propagation simulation results of step S3, combined with mechanical parameters such as the crack growth rate and crack tip stress intensity factor of transparent coal rock, to perform time-series modeling of the crack initiation and expansion process in a three-dimensional dynamic visualization observation model. The crack growth rate reflects the speed at which the crack develops over time, while the crack tip stress intensity factor is a key parameter for measuring the strength of the stress field at the crack tip. By incorporating these mechanical parameters into the modeling process, the morphological changes of the crack at different time nodes can be simulated, and a three-dimensional model sequence containing the crack morphology at different time nodes can be generated. This time-series modeling method can dynamically display the entire process of cracks from initial initiation to gradual expansion, revealing the laws of crack evolution over time within coal rock, helping researchers to deeply understand the mechanical mechanism of crack growth and providing an important basis for predicting crack development trends.

[0040] During implementation, the areas where cracks may exist and develop are determined based on the results of light propagation simulation. Combined with mechanical parameters such as the transparent coal rock crack growth rate and crack tip stress intensity factor obtained through experimental measurement or theoretical calculation, a specific modeling algorithm is used to model the cracks in a three-dimensional dynamic visualization observation model. In chronological order, the crack growth within each time interval is calculated step by step, the crack morphology and position are updated, and a series of three-dimensional models at different time nodes are generated. Each three-dimensional model accurately reflects the actual state of the crack inside the coal rock at the corresponding moment. By arranging these models in chronological order, a complete crack dynamic evolution sequence is formed, which enables the crack growth process to be presented in an intuitive three-dimensional dynamic form, providing researchers with a comprehensive and in-depth perspective for observing crack development.

[0041] Step S5: Using the non-maximum suppression and bounding box regression mechanism in the optimized SSD algorithm, the crack area in the 3D model sequence is accurately located and contour extracted to obtain the geometric size and spatial position parameters of the crack;

[0042] Specifically, step S5 uses the non-maximum suppression and bounding box regression mechanisms in the optimized SSD algorithm to accurately locate and extract the contours of the crack areas in the 3D model sequence generated in step S4, obtaining the geometric dimensions and spatial position parameters of the cracks. The non-maximum suppression mechanism can remove repeated and redundant detection results and retain the most accurate crack location information; the bounding box regression mechanism accurately adjusts the boundaries of the crack area to make the determined crack contour more consistent with the actual form. When processing a sequence of transparent coal and rock 3D models, the algorithm analyzes the potential crack areas in each model, selects the optimal crack detection results through non-maximum suppression, and then uses bounding box regression to refine the crack boundaries, thereby accurately obtaining the geometric dimensions of the crack, such as length and width, as well as the specific position coordinates in 3D space, providing accurate data for subsequent quantitative analysis and research on the cracks.

[0043] During the implementation process, the 3D model sequence is input into the optimized SSD algorithm. The algorithm first uses the non-maximum suppression mechanism to evaluate and screen the multiple crack candidate regions detected in the model. By comparing the confidence and overlap of different candidate regions, the regions with high confidence and no overlap or low overlap are retained as the final crack detection results. Then, using the bounding box regression mechanism, the bounding box of the detection result is fine-tuned according to the actual characteristics of the crack area, so that the bounding box can more accurately enclose the actual contour of the crack. Through the coordinated processing of these two mechanisms, the key parameters of the crack can be accurately extracted from the complex 3D model sequence, providing reliable data support for in-depth research on the characteristics and evolution of cracks, and improving the accuracy and effectiveness of crack analysis.

[0044] Step S6: The geometric dimensions, spatial position parameters and three-dimensional models of crack morphology at different time nodes are integrated and presented in a three-dimensional dynamic visualization observation model to form a dynamic visualization observation result of cracks inside transparent coal rock.

[0045] Specifically, step S6 integrates the crack geometry and spatial position parameters obtained in step S5 and the three-dimensional model of crack morphology at different time nodes generated in step S4 into a three-dimensional dynamic visualization observation model to form a dynamic visualization observation result of cracks inside transparent coal rock. This step integrates the key information obtained in the previous steps, combines the static parameters of the crack (geometric dimensions and spatial position) with the dynamic evolution process (morphology at different time nodes), and displays them in an intuitive and comprehensive manner in the three-dimensional dynamic visualization observation model. Through this fusion presentation, researchers can simultaneously observe the specific morphology and size of the cracks and their changes over time, which facilitates the analysis and study of cracks from multiple angles, and provides a clear and intuitive visualization tool for in-depth understanding of the initiation and expansion mechanism of cracks inside transparent coal rock.

[0046] During the implementation process, the geometric dimensions, spatial position parameters, and three-dimensional model data of the crack at different time nodes are first imported into the three-dimensional dynamic visualization observation model. Based on this data, the model accurately draws the shape of the crack at each time node in three-dimensional space and marks its geometric dimensions and spatial position information. By setting appropriate visualization parameters, such as lighting, viewing angle, color coding, etc., the dynamic evolution process of the crack is displayed in a smooth and clear animation form. Through interactive operations, researchers can observe the details and overall change trends of the crack from different angles, and realize a full-scale, multi-level observation of the dynamic evolution process of cracks inside transparent coal rocks. This visual presentation method greatly improves the efficiency and accuracy of researchers' crack research and helps promote research progress in related fields.

[0047] Preferably, in step S2, the optimized SSD algorithm adopts an improved feature pyramid structure to construct the following feature fusion model:

[0048]

[0049] Among them, F merged is the fused feature map; are feature maps of different levels respectively; α, β, and γ are fusion weights, whose values ​​are determined by an adaptive learning algorithm based on physical parameters such as the elastic modulus and Poisson's ratio of transparent coal rock, and are used to balance the contribution of feature maps of different levels to crack feature extraction.

[0050] Specifically, during the feature extraction process in step S2, the feature pyramid structure of the SSD algorithm is improved to more accurately capture information about potential crack regions within discrete units of transparent coal rock. A feature fusion mechanism is established to integrate feature maps extracted at different levels. The determination of fusion weights is closely related to physical parameters of the transparent coal rock, such as its elastic modulus and Poisson's ratio. Using an adaptive learning algorithm, the weights are dynamically adjusted based on the contribution of these parameters to crack feature extraction. This allows the detailed features captured at the bottom layer and the semantic features extracted at the higher layer to complement each other, effectively improving the accuracy and comprehensiveness of crack feature extraction.

[0051] Preferably, in step S3, the propagation path of light inside the transparent coal rock is determined using the following model: in, is the position vector of the light in the nth discrete unit body; is the position vector of the light in the n+1th discrete unit body; Δs is the distance that the light propagates between adjacent discrete units, which is determined according to the density, transmittance and other parameters of the transparent coal rock; is the propagation direction vector of the light in the nth discrete unit body, which is determined by the refractive index of the transparent coal rock and the geometric shape of the crack surface.

[0052] Specifically, step S3 determines the propagation path of light within the transparent coal rock by establishing a computational relationship based on the optical and structural properties of the transparent coal rock. As light propagates between discrete units of the coal rock, its propagation distance is determined by parameters such as the density and transmittance of the transparent coal rock, while its propagation direction vector depends on the refractive index of the coal rock and the geometry of the crack surface. By clarifying the recursive relationship between the position vectors of light between adjacent discrete units, the entire process of light entering the transparent coal rock sample and being refracted, reflected, and scattered by the cracks can be accurately simulated, providing a reliable data foundation for optical analysis of crack location and morphology.

[0053] Preferably, in step S4, the time series modeling of the crack propagation process adopts the following dynamic evolution model:

[0054]

[0055] Where C(t) is the crack profile at time t; C(t+Δt) is the crack profile at time t+Δt; δ is the crack growth increment coefficient, which is determined according to the crack growth rate of transparent coal rock; K I (t) is the stress intensity factor at the crack tip at time t; K IC is the fracture toughness of transparent coal rock; is the normal vector of the crack tip at time t.

[0056] Specifically, in step S4, when modeling the crack propagation process in a time series, a dynamic evolution model is constructed in conjunction with the mechanical property parameters of the transparent coal rock. Based on the crack profile at different times, the expansion increment coefficient is determined according to the crack propagation rate of the transparent coal rock. The change in the crack profile at the next moment is calculated by combining the stress intensity factor at the crack tip with the material's fracture toughness and the crack tip normal vector. This method, combining time variables with mechanical parameters, systematically simulates the dynamic process of crack initiation and propagation, generating a sequence of three-dimensional crack morphology models at different time points, intuitively demonstrating the crack propagation pattern.

[0057] Preferably, in step S5, the non-maximum suppression mechanism used for crack area positioning introduces the grayscale distribution parameters of the transparent coal rock for optimization, and the optimized non-maximum suppression threshold determination model is as follows: T = μ + λ·σ, wherein T is the non-maximum suppression threshold; μ is the mean of the grayscale value of the transparent coal rock; σ is the standard deviation of the grayscale value of the transparent coal rock; λ is the adjustment coefficient, which is determined according to the transmittance of the transparent coal rock and the contrast between the crack and the background.

[0058] Specifically, in step S5, when locating crack regions, an optimized threshold determination method is introduced using the grayscale distribution parameters of transparent coal rock to improve the accuracy of the non-maximum suppression mechanism. The mean and standard deviation of the transparent coal rock grayscale values ​​reflect the overall grayscale characteristics of the sample, while the adjustment coefficient is determined based on the coal rock transmittance and the contrast between the crack and the background. By incorporating these parameters into the threshold calculation, the algorithm can adaptively adjust the non-maximum suppression criteria based on the characteristics of the transparent coal rock sample, effectively removing redundant detection results, accurately locating the crack region, and accurately extracting the crack outline using a bounding box regression mechanism.

[0059] Preferably, in step S1, the spatial gridding process of the transparent coal rock sample is combined with the particle size distribution parameters of the transparent coal rock, and a variable resolution gridding model is adopted, specifically as follows: N cell =f(d particle ,ρ density ), where N cell is the number of units after meshing; d particle is the average particle size of transparent coal rock particles; ρ density is the density of transparent coal rock; f is the mapping function established according to the physical properties of transparent coal rock, which is used to determine the grid resolution of different areas.

[0060] Specifically, in step S1, when spatially gridding the transparent coal rock sample, a variable-resolution gridding strategy is employed, taking into account the heterogeneity of the coal rock's internal structure and combining its particle size distribution parameters. The number of gridded units is related to the average particle size and density of the transparent coal rock particles. A mapping function is established to determine the grid resolution for different regions based on these physical properties. A high-resolution grid is used in regions with small particle size and complex structure to capture crack details. The resolution is reduced in regions with relatively simple structures, balancing computational accuracy and efficiency to achieve efficient and accurate gridding of the sample.

[0061] Preferably, in step S2, in the convolutional neural network layer of the optimized SSD algorithm, the convolution kernel parameters are adjusted according to the texture characteristic parameters of the transparent coal rock, using the following adaptive adjustment model: Among them, w ij is the adjusted convolution kernel parameter; is the initial convolution kernel parameter; Δw is the parameter adjustment step; g(θ texture ) is the texture direction of transparent coal rock θ texture Related functions are used to enable the convolution kernel to better extract crack texture features.

[0062] Specifically, in the SSD algorithm convolutional neural network layer optimized in step S2, to enhance the ability to extract transparent coal crack texture features, the convolution kernel parameters are adaptively adjusted based on the coal texture feature parameters. Based on the initial convolution kernel parameters, the parameter adjustment step size is determined based on the transparent coal texture direction. A specific function is used to convert the texture direction features into a basis for convolution kernel parameter adjustment. This allows the convolution kernel to better match the crack texture features during the calculation process, thereby improving the algorithm's targeted and effective crack feature extraction.

[0063] Preferably, in step S4, the rendering of the crack morphology in the three-dimensional dynamic visualization observation model adopts an illumination model based on the optical parameters of transparent coal rock, specifically: Where, I is the final illumination intensity on the crack surface; k a 、k d 、k s are the ambient light reflection coefficient, diffuse reflection coefficient, and specular reflection coefficient, respectively, which are determined according to the transmittance and refractive index of transparent coal rock; I a , I d , I s They are ambient light intensity, diffuse light intensity, and specular light intensity respectively; is the crack surface normal vector; is the light direction vector; is the viewpoint direction vector; is the direction vector of the reflected light; n is the glossiness index, which is determined according to the surface roughness of the transparent coal rock.

[0064] Specifically, in step S4, when rendering the crack morphology in the 3D dynamic visualization observation model, an illumination model is constructed based on the optical parameters of the transparent coal rock. The ambient light reflection coefficient, diffuse reflection coefficient, and specular reflection coefficient are determined by the transmittance and refractive index of the transparent coal rock, while the gloss index depends on the surface roughness of the coal rock. By comprehensively considering different light components such as ambient light, diffuse reflection, and specular reflection, as well as geometric relationships such as the crack surface normal vector and the light direction vector, the final illumination intensity of the crack surface is calculated, realistically reproducing the visual effects of the crack under different lighting conditions, improving the authenticity and intuitiveness of the visualization results.

[0065] Preferably, in step S6, the fusion presentation of the three-dimensional dynamic visualization observation results adopts a multimodal data fusion model, which combines the geometric parameters and mechanical parameters of the transparent coal rock as follows: final =ω1·V geometry +ω2·V mechanics , where V final is the final three-dimensional dynamic visualization observation result; V geometry V is a visualization model based on the crack geometry and spatial position; mechanics is a visualization model based on the mechanical properties of cracks; ω1 and ω2 are fusion weights, which are determined by the weighted average algorithm according to the elastic modulus, crack growth rate and other parameters of transparent coal rock.

[0066] Specifically, step S6 employs a multimodal data fusion model to fuse and present the 3D dynamic visualization observation results. A visualization model based on the crack's geometric dimensions and spatial position is integrated with a visualization model based on the crack's mechanical properties. The fusion weights are determined using a weighted average algorithm based on parameters such as the elastic modulus and crack growth rate of the transparent coal. This fusion approach combines the crack's geometric information with its mechanical evolution information, enabling researchers to gain a comprehensive understanding of crack characteristics from multiple dimensions and providing a richer data presentation for in-depth analysis.

[0067] like Figure 2 As shown in FIG, a dynamic visualization observation method of transparent coal rock materials and internal cracks is implemented through different units, including:

[0068] A sample space construction unit is used to construct a three-dimensional spatial coordinate system of the transparent coal rock material and perform spatial grid processing on the transparent coal rock sample based on the physical parameters of the transparent coal rock;

[0069] The feature extraction processing unit is connected to the sample space construction unit and is used to extract multi-scale features of the transparent coal rock discrete unit body using the optimized SSD algorithm to obtain a feature map containing information about potential crack areas;

[0070] The light propagation simulation unit is connected to the feature extraction processing unit, and performs light propagation simulation on the feature map based on the optical parameters of the transparent coal rock to determine the propagation path of light generated by cracks inside the transparent coal rock;

[0071] The crack evolution modeling unit is connected to the light propagation simulation unit. Combined with the mechanical parameters of transparent coal rock, it performs time-series modeling of the crack initiation and propagation process in a three-dimensional dynamic visualization observation model.

[0072] The crack location and extraction unit is connected to the crack evolution modeling unit. It accurately locates and extracts the contours of the crack area through the relevant mechanisms in the optimized SSD algorithm.

[0073] The result fusion presentation unit is connected to the crack location extraction unit, which fuses the crack-related parameters with the three-dimensional models of crack morphology at different time nodes in a three-dimensional dynamic visualization observation model to form the final observation results.

[0074] Existing coal and rock internal crack observation technologies have limitations in accuracy and dynamic presentation. This method relies on an optimized SSD algorithm and a three-dimensional dynamic visualization model, and deeply combines transparent coal and rock parameters to effectively break through these bottlenecks and demonstrate significant advantages.

[0075] To address the shortcomings of traditional methods in being unable to comprehensively utilize the material properties of transparent coal rock, this observation method constructs a multi-parameter fusion system. In the spatial modeling stage, the sample is divided into three-dimensional grids based on physical parameters such as elastic modulus and Poisson's ratio, laying a precise spatial foundation for subsequent analysis. In the light simulation stage, optical parameters such as transmittance and refractive index are used to accurately simulate the refraction and reflection paths of light in cracked transparent coal rock. Compared with the traditional single-parameter judgment mode, it can more accurately reflect the true shape and location of the cracks. At the same time, the optimized SSD algorithm uses a feature fusion model to adaptively adjust the weights according to the physical properties of transparent coal rock, achieving efficient extraction of multi-scale crack features, greatly improving the accuracy and reliability of crack identification.

[0076] In terms of overcoming the lack of dynamic observation capabilities, this method achieves a technological breakthrough through multi-step collaboration. Based on mechanical parameters such as crack propagation rate and stress intensity factor, the crack initiation and propagation process is modeled in time series, and a three-dimensional model sequence at different time nodes is generated to quantitatively present the dynamic evolution process of the crack. With the help of the improved non-maximum suppression and bounding box regression mechanism in the optimized SSD algorithm, combined with the threshold model based on the optimization of the grayscale distribution parameters of transparent coal rock, accurate positioning and contour extraction of the crack area are achieved, and subtle changes in the crack can be captured in real time. In addition, the three-dimensional dynamic visualization observation model uses an illumination model based on optical parameters and a multimodal data fusion model to enhance the display effect of the crack dynamic information from a visualization level, making the observation results more intuitive and comprehensive.

[0077] In summary, this method, based on the dynamic visualization observation of transparent coal and rock materials and their internal cracks, forms a complete observation system by deeply integrating the multi-dimensional parameters of transparent coal and rock, and organically integrating algorithms and models. This method optimizes the entire process, from spatial construction, feature extraction, and light simulation to dynamic modeling and result presentation. This method not only effectively overcomes the shortcomings of existing technologies but also achieves a qualitative leap in observation accuracy, dynamic capture capabilities, and visualization effects. It provides an advanced and reliable technical means for studying internal cracks in transparent coal and rock, and has important practical significance and academic value for promoting the development of fields such as coal mining and geomechanics.

[0078] In the description of the present invention, it should be noted that, unless otherwise expressly specified or limited, the terms "disposed," "installed," "connected," "connected," and "fixed" should be understood in a broad sense. For example, they may refer to fixed connections, detachable connections, or integral connections; they may refer to mechanical connections or electrical connections; they may refer to direct connections or indirect connections through an intermediate medium; and they may refer to internal communication between two components. Those skilled in the art will understand the specific meanings of the above terms in the present invention based on specific circumstances.

[0079] While embodiments of the present invention have been shown and described, it will be understood by those skilled in the art that various equivalent changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A method for dynamic visualization observation of transparent coal rock materials and internal cracks thereof, characterized in that: The method includes: Step S1: constructing a three-dimensional spatial coordinate system for the transparent coal rock material, performing spatial gridding processing on the transparent coal rock sample based on the elastic modulus, Poisson's ratio, and density physical parameters of the transparent coal rock, and dividing the transparent coal rock sample into discrete units with three-dimensional spatial coordinates; Step S2: Using the optimized SSD algorithm, a multi-scale feature extraction mechanism is established for the surface features and internal structure of discrete units of transparent coal rock. Through feature extraction and fusion of convolutional neural network layers, a feature map containing information about potential crack areas is obtained. Step S3: Based on the optical parameters of the transparent coal rock, such as transmittance and refractive index, a light propagation simulation is performed on the characteristic graph in a three-dimensional spatial coordinate system to construct an interaction relationship between the light and the discrete units of the transparent coal rock, thereby determining the refraction, reflection, and scattering paths caused by cracks when the light propagates inside the transparent coal rock; Step S4: Based on the light propagation simulation results, combined with the crack growth rate and crack tip stress intensity factor mechanical parameters of the transparent coal rock, the crack initiation and growth process are modeled in a three-dimensional dynamic visualization observation model to generate a three-dimensional model sequence containing crack morphologies at different time nodes; Step S5: Using the non-maximum suppression and bounding box regression mechanism in the optimized SSD algorithm, the crack area in the 3D model sequence is accurately located and contour extracted to obtain the geometric size and spatial position parameters of the crack; Step S6: The geometric dimensions, spatial position parameters and three-dimensional models of crack morphology at different time nodes are integrated and presented in a three-dimensional dynamic visualization observation model to form a dynamic visualization observation result of cracks inside transparent coal rock.

2. The method for dynamic visualization observation of transparent coal rock materials and internal cracks according to claim 1 is characterized in that: In step S2, the optimized SSD algorithm adopts an improved feature pyramid structure to construct the following feature fusion model: Among them, F merged is the fused feature map; are feature maps of different levels respectively; α, β, and γ are fusion weights, whose values ​​are determined by the adaptive learning algorithm based on the elastic modulus and Poisson's ratio physical parameters of transparent coal rock, and are used to balance the contribution of feature maps of different levels to crack feature extraction.

3. The method for dynamic visualization observation of transparent coal rock materials and internal cracks thereof according to claim 1 is characterized in that: In step S3, the propagation path of light inside the transparent coal rock is determined using the following model: in, is the position vector of the light in the nth discrete unit body; is the position vector of the light in the n+1th discrete unit body; Δs is the distance that the light propagates between adjacent discrete units, which is determined according to the density and transmittance parameters of the transparent coal rock; is the propagation direction vector of the light in the nth discrete unit body, which is determined by the refractive index of the transparent coal rock and the geometric shape of the crack surface.

4. The method for dynamic visualization observation of transparent coal rock materials and internal cracks according to claim 1 is characterized in that: In step S4, the time series modeling of the crack propagation process adopts the following dynamic evolution model: Where C(t) is the crack profile at time t; C(t+Δt) is the crack profile at time t+Δt; δ is the crack growth increment coefficient, which is determined according to the crack growth rate of transparent coal rock; K I (t) is the stress intensity factor at the crack tip at time t; K IC is the fracture toughness of transparent coal rock; is the normal vector of the crack tip at time t.

5. The method for dynamic visualization observation of transparent coal rock materials and internal cracks thereof according to claim 1 is characterized in that: In step S5, the non-maximum suppression mechanism used for crack area positioning introduces the grayscale distribution parameters of the transparent coal rock for optimization. The optimized non-maximum suppression threshold determination model is as follows: T = μ + λ·σ, where T is the non-maximum suppression threshold; μ is the mean of the grayscale value of the transparent coal rock; σ is the standard deviation of the grayscale value of the transparent coal rock; λ is the adjustment coefficient, which is determined according to the transmittance of the transparent coal rock and the contrast between the crack and the background.

6. The method for dynamic visualization observation of transparent coal rock materials and internal cracks according to claim 1 is characterized in that: In step S1, the spatial gridding process of the transparent coal rock sample is combined with the particle size distribution parameters of the transparent coal rock, and a variable resolution gridding model is used, specifically as follows: N cell =f(d particle , ρ density ), where N cell is the number of units after meshing; d particle is the average particle size of transparent coal rock particles; ρ density is the density of transparent coal rock; f is the mapping function established according to the physical properties of transparent coal rock, which is used to determine the grid resolution of different areas.

7. The method for dynamic visualization observation of transparent coal rock materials and internal cracks thereof according to claim 1 is characterized in that: In step S2, in the convolutional neural network layer of the optimized SSD algorithm, the convolution kernel parameters are adjusted according to the texture feature parameters of the transparent coal rock, using the following adaptive adjustment model: Among them, w ij is the adjusted convolution kernel parameter; is the initial convolution kernel parameter; Δw is the parameter adjustment step; g(θ texture ) is the texture direction of transparent coal rock θ texture Related functions are used to enable the convolution kernel to better extract crack texture features.

8. The method for dynamic visualization observation of transparent coal rock materials and internal cracks thereof according to claim 1 is characterized in that: In step S4, the rendering of the crack morphology in the three-dimensional dynamic visualization observation model adopts an illumination model based on the optical parameters of transparent coal rock, specifically: Where, I is the final illumination intensity on the crack surface; k a 、k d 、k s are the ambient light reflection coefficient, diffuse reflection coefficient, and specular reflection coefficient, respectively, which are determined according to the transmittance and refractive index of transparent coal rock; I a , I d , I s They are ambient light intensity, diffuse light intensity, and specular light intensity respectively; is the crack surface normal vector; is the light direction vector; is the viewpoint direction vector; is the direction vector of the reflected light; n is the glossiness index, which is determined according to the surface roughness of the transparent coal rock.

9. The method for dynamic visualization observation of transparent coal rock materials and internal cracks thereof according to claim 1, characterized in that: In step S6, the fusion presentation of the three-dimensional dynamic visualization observation results adopts a multimodal data fusion model, which combines the geometric parameters and mechanical parameters of the transparent coal rock as follows: final =ω1·V geometry +ω2·V mechanics , where V final is the final three-dimensional dynamic visualization observation result; V geometry V is a visualization model based on the crack geometry and spatial position; mechanics is a visualization model based on the mechanical properties of cracks; ω1 and ω2 are fusion weights, which are determined by the weighted average algorithm according to the elastic modulus and crack growth rate parameters of transparent coal rock.

Images