An intelligent identification method and system for mine ecological damage based on remote sensing and support vector machine

Abstract

The application provides a mine ecological damage intelligent identification method and system based on remote sensing and support vector machines, relating to the technical field of mine ecological environment monitoring. The method includes collecting multi-source remote sensing data to construct an initial data set, extracting comprehensive ecological damage features through a three-dimensional residual convolutional neural network, and realizing nonlinear classification using a hybrid kernel function support vector machine. At the same time, the ecological damage index is calculated and a distribution map is generated, combined with a dynamic threshold to divide the damage level and trigger a three-level early warning mechanism to generate repair suggestions. The method enhances the expression ability of ecological features by introducing a deep learning model and multi-source remote sensing data fusion technology. Through the support vector machine algorithm and kernel function mapping, the adaptability of the classification model in complex scenarios is improved. Through the dynamic threshold and multi-level early warning mechanism, accurate grading and timely response to ecological damage are realized, providing a scientific basis and technical support for mine ecological environment monitoring and repair.

Description

Technical Field

[0001] This invention relates to the field of mine ecological environment monitoring technology, and in particular to an intelligent identification method and system for mine ecological damage based on remote sensing and support vector machine. Background Technology

[0002] With the rapid development of remote sensing technology and machine learning algorithms, ecological damage identification methods based on remote sensing imagery and intelligent algorithms have gradually become a research hotspot in the field of mine ecological environment monitoring. During mining operations, rapid identification and assessment of ecological damage are crucial for achieving sustainable development of the mining area environment. However, existing mine ecological damage identification methods still have many shortcomings in data processing, feature extraction, and classification model construction, affecting identification accuracy and application efficiency.

[0003] Currently, invention patent CN113744249B discloses a method for investigating marine ecological environment damage. This method constructs a remote sensing survey information database for marine ecological environment damage, acquires data by combining optimal remote sensing observation methods, and generates damage survey results using image preprocessing and feature extraction techniques. However, the feature extraction of remote sensing images in this method mainly relies on traditional image processing methods, failing to fully utilize the advantages of advanced classification algorithms such as Support Vector Machines (SVM), which leads to insufficient accuracy in identifying ecological damage in complex scenarios. Therefore, it is essential to design an intelligent identification method and system for mine ecological damage based on remote sensing and support vector machines. Summary of the Invention

[0004] The purpose of this invention is to provide a method and system for intelligent identification of mine ecological damage based on remote sensing and support vector machine, so as to improve the adaptability of classification models in complex scenarios.

[0005] To achieve the above objectives, the present invention provides the following solution:

[0006] A method for intelligent identification of mine ecological damage based on remote sensing and support vector machine includes the following steps:

[0007] In the mining area and surrounding ecologically sensitive areas, multi-source remote sensing data including high-resolution satellite imagery, multispectral imagery and radar imagery were collected simultaneously to obtain an initial dataset.

[0008] The initial dataset is preprocessed, and ecological features including spatial, spectral, and textural features are extracted using a three-dimensional convolutional neural network. These ecological features are then fused with topographic, geological, and meteorological auxiliary data to obtain an ecological damage feature set.

[0009] Based on the ecological damage feature set, a mine ecological damage classification model is constructed using the support vector machine algorithm. The nonlinear classification problem of the mine ecological damage classification model is solved by kernel function mapping to obtain the ecological damage type and spatial distribution.

[0010] The ecological damage index of the mining area is calculated based on the type of ecological damage, and the ecological damage value is obtained by combining spatial analysis algorithm. The ecological damage value is then normalized.

[0011] Ecological damage levels are classified based on normalized ecological damage values. When the ecological damage value reaches or exceeds a preset threshold, a multi-level early warning mechanism is triggered and ecological restoration suggestions are generated.

[0012] Optionally, multi-source remote sensing data, including high-resolution satellite imagery, multispectral imagery, and radar imagery, are simultaneously collected in the mining area and surrounding ecologically sensitive areas to obtain an initial dataset, including:

[0013] Based on the mining area and topographic diffusion model, data collection points were determined in the core mining area, transition area and ecological protection area respectively;

[0014] High-resolution satellite images of data collection points are acquired through a satellite remote sensing platform, and information on land cover change is extracted.

[0015] Multispectral images of data collection points were collected using a drone-borne hyperspectral instrument, and vegetation stress index, water body index, and building index were calculated.

[0016] Surface deformation and soil moisture were extracted from radar images;

[0017] The initial dataset was obtained by spatiotemporally aligning the information on land cover change, vegetation stress index, water body index, building index, land surface deformation, and soil moisture.

[0018] Optionally, the initial dataset is preprocessed, and ecological features including spatial, spectral, and textural features are extracted using a 3D convolutional neural network. These ecological features are then fused with topographic, geological, and meteorological auxiliary data to obtain an ecological damage feature set, including:

[0019] Transformation vector analysis was performed on the initial dataset to extract key time-phase nodes of the time series;

[0020] Construct a three-dimensional convolutional neural network that includes spatial convolutional layers, spectral convolutional layers, texture convolutional layers, and feature fusion layers;

[0021] Based on the surface deformation gradient of the mining area extracted from the key temporal nodes of the spatial convolution layer, spatial features including the boundary vector of the mining face, the expansion area of ​​the spoil heap, and the surface subsidence gradient are obtained.

[0022] The spectral features, including vegetation red edge position offset, mineral characteristic absorption depth and water turbidity index, are obtained from the spectral derivatives of key temporal nodes extracted by the spectral convolution layer.

[0023] Texture features, including surface fragmentation, fracture network density, and vegetation distribution heterogeneity index, are extracted from key temporal nodes based on texture convolutional layers.

[0024] Ecological features are obtained by splicing and dimensionality reduction of spatial, spectral, and texture features.

[0025] Optionally, a change vector analysis is performed on the initial dataset to extract key time-phase nodes of the time series, including:

[0026] The difference operation is performed on any two data points from different time phases in the initial dataset in the multidimensional spectral space to obtain the change vector;

[0027] Calculate the magnitude and orientation angle of the change vector; the magnitude is used to represent the drasticness of the data change, and the orientation angle is used to represent the spectral type of the data change;

[0028] Nodes with local peak values ​​in amplitude are identified, and nodes with the same orientation angle as the preset orientation angle are selected as key time phase nodes.

[0029] Optionally, based on the ecological damage feature set, a mine ecological damage classification model is constructed using the support vector machine algorithm. The nonlinear classification problem of the mine ecological damage classification model is then solved using kernel function mapping to obtain the ecological damage types and spatial distribution, including:

[0030] A historical dataset is generated based on historical image data of the mining area and surrounding ecologically sensitive areas during the period before mining operations began.

[0031] The Gaussian radial basis function kernel is used as the kernel function of the support vector machine model, and the parameters of the support vector machine model are optimized by maximizing the classification margin to obtain the ecological damage classification model.

[0032] The correlation between the ecological damage feature set and the historical dataset is calculated using an ecological damage classification model, and the types and spatial distribution of ecological damage are obtained based on the correlation.

[0033] Optionally, an ecological damage index for the mining area is calculated based on the type of ecological damage, and an ecological damage value is obtained by combining it with a spatial analysis algorithm. The ecological damage value is then normalized, including:

[0034] Based on the type of ecological damage, the ecological damage index for each spatial grid unit in the mining area is calculated; the formula for calculating the ecological damage index is: ,in, These are the weighting coefficients. This represents the probability of type c damage.

[0035] The ecological damage index was corrected by spatial autocorrelation analysis and kernel density estimation to obtain the corrected damage index;

[0036] The ecological damage value is calculated based on the modified damage index and environmental sensitivity factors; the formula for calculating the ecological damage value is: ,in, To correct the damage index, This is the adjustment coefficient for the sensitive area. Sensitive area level, For the current grid rainfall erosion force, The regional benchmark erosion force.

[0037] The ecological damage values ​​were normalized using the maximum-minimum normalization method.

[0038] Optionally, the ecological damage index is corrected through spatial autocorrelation analysis and kernel density estimation to obtain a corrected damage index, including:

[0039] The spatial autocorrelation index is calculated based on the ecological damage index; the formula for calculating the spatial autocorrelation index is: ,in, This is the spatial weight matrix. n The number of neighboring grid cells. This represents the average value of the ecological damage index;

[0040] Damage diffusion density of spatial autocorrelation index is obtained by using Gaussian kernel function;

[0041] The modified damage index is obtained based on the damage diffusion density and spatial autocorrelation index; the formula for calculating the modified damage index is: ,in This represents the damage diffusion density.

[0042] Optionally, ecological damage levels are classified based on normalized ecological damage values. When an ecological damage value reaches or exceeds a preset threshold, a multi-level early warning mechanism is triggered, and ecological restoration suggestions are generated, including:

[0043] When the ecological damage value is within [0,3), the ecological damage level is determined to be mild; when the ecological damage value is within [3,6), the ecological damage level is determined to be moderate; when the ecological damage value is within [6,8), the ecological damage level is determined to be severe; when the ecological damage value is greater than or equal to 8, the ecological damage level is determined to be extremely severe.

[0044] A dynamic threshold is generated based on environmental parameters, including rainfall erosivity index and topographic relief. The formula for calculating the dynamic threshold is: Where TI is the topographic relief and RI is the rainfall erosivity index;

[0045] When the ecological damage value is ≥ A blue alert is generated when the ecological damage value is ≥ max( A yellow alert is generated when the ecological damage value is ≥6 (3.5); an orange alert is generated when the ecological damage value is ≥8 (3.5); and an orange alert is generated when the ecological damage value is ≥8 (3.5). A red alert is generated when the value is greater than 0.5.

[0046] A smart identification system for mine ecological damage based on remote sensing and support vector machine includes:

[0047] The data acquisition module is used to simultaneously collect multi-source remote sensing data, including high-resolution satellite imagery, multispectral imagery, and radar imagery, in the mining area and surrounding ecologically sensitive areas to obtain an initial dataset.

[0048] The data preprocessing module is used to preprocess the initial dataset, extract ecological features including spatial, spectral, and textural features through a three-dimensional convolutional neural network, and fuse the ecological features with topographic, geological, and meteorological auxiliary data to obtain an ecological damage feature set.

[0049] The classification modeling module is used to construct a mine ecological damage classification model based on the ecological damage feature set using the support vector machine algorithm, and to solve the nonlinear classification problem of the mine ecological damage classification model through kernel function mapping, so as to obtain the ecological damage type and spatial distribution.

[0050] The index calculation module is used to calculate the ecological damage index of the mining area according to the type of ecological damage, and to obtain the ecological damage value by combining spatial analysis algorithm, and to normalize the ecological damage value;

[0051] The early warning generation module is used to classify the ecological damage level based on the normalized ecological damage value. When the ecological damage value reaches or exceeds the preset threshold, a multi-level early warning mechanism is triggered and ecological restoration suggestions are generated.

[0052] According to specific embodiments provided by the present invention, the following technical effects are disclosed: The intelligent identification method for mine ecological damage based on remote sensing and support vector machine provided by the present invention includes: simultaneously collecting multi-source remote sensing data including high-resolution satellite imagery, multispectral imagery, and radar imagery in the mining area and surrounding ecologically sensitive areas to obtain an initial dataset; preprocessing the initial dataset, extracting ecological features including spatial features, spectral features, and texture features through a three-dimensional convolutional neural network, and fusing the ecological features with topographic, geological, and meteorological auxiliary data to obtain an ecological damage feature set; constructing a mine ecological damage classification model based on the ecological damage feature set using a support vector machine algorithm, and solving the nonlinear classification problem of the mine ecological damage classification model through kernel function mapping to obtain the ecological damage type and spatial distribution; calculating the ecological damage index of the mining area according to the ecological damage type, and obtaining the ecological damage value by combining spatial analysis algorithm, and normalizing the ecological damage value; classifying the ecological damage level according to the normalized ecological damage value, and triggering a multi-level early warning mechanism and generating ecological restoration suggestions when the ecological damage value reaches or exceeds a preset threshold. This method enhances the expressive power of ecological features by introducing deep learning models and multi-source remote sensing data fusion technology; improves the adaptability of classification models in complex scenarios through support vector machine algorithms and kernel function mapping; and achieves accurate classification and timely response to ecological damage through dynamic thresholds and multi-level early warning mechanisms, providing scientific basis and technical support for mine ecological environment monitoring and restoration. Attached Figure Description

[0053] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0054] Figure 1 This is a flowchart of the intelligent identification method for mine ecological damage according to the present invention;

[0055] Figure 2 The flowchart for constructing the ecological damage feature set of this invention is shown below. Detailed Implementation

[0056] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0057] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0058] like Figure 1 As shown, this invention provides an intelligent identification method for mine ecological damage based on remote sensing and support vector machines, comprising the following steps:

[0059] Step 100: In the mining area and surrounding ecologically sensitive areas, simultaneously collect multi-source remote sensing data including high-resolution satellite imagery, multispectral imagery and radar imagery to obtain an initial dataset;

[0060] Specifically, this embodiment first calculates the impact range based on the coordinates of the mine's mining center and the digital elevation model (DEM). A circular area with a radius of 0.5 times the permitted mining area is designated as the core mining zone, centered on the mining point; then, based on the topographic diffusion equation... Determine the boundaries of the transition zone, where K The diffusion coefficient is 1.2 in this embodiment. α The terrain attenuation factor is 0.5 in this embodiment. The elevation gradient was used; simultaneously, the boundary of the ecological red line protection zone was extended by 200 meters to form an ecological protection zone. After the division, data collection points were deployed in a 1km×1km grid within the three areas, ensuring that each grid contained at least one collection point.

[0061] Then, high-resolution satellite imagery of the data collection points was acquired via a satellite remote sensing platform. The imagery was segmented using the object-oriented classification method of the eCognition platform, with a segmentation scale parameter of 50, a shape factor of 0.3, and a compactness of 0.7. The Normalized Difference Vegetation Index (NDVI) was then calculated for the two temporal images using the following formula: In this context, NIR represents the near-infrared band, and Red represents the red band. Areas with an NDVI difference greater than 0.2 between two time phases are marked as land cover change areas, and the NDVI difference reflects land cover change information. Multispectral images of the data collection points were acquired using an UAV-borne hyperspectral imager, and vegetation stress index, water index, and building index were calculated. The vegetation stress index is the sum of red edge offset (REIP) and chlorophyll fluorescence index (SIF), calculated using the following formulas: , ,in Let λ be the reflectance at wavelength λ. This is a reference value for undisturbed areas. The formula for calculating the Water Index (TSI) is: Where Green represents the green band. The formula for calculating the Building Index is: ,in, Pixel values ​​in the Gray-Level Co-occurrence Matrix (GLCM)i The probability of obtaining the data was determined. Simultaneously, surface deformation and soil moisture in the same area were extracted from the radar images. Finally, bilinear interpolation was used to convert the coordinate systems in all images to UTM projection to ensure grid center point alignment. The six data categories—land cover change information, vegetation stress index, water index, building index, surface deformation, and soil moisture—were aggregated into a multidimensional matrix according to the grid, resulting in the initial dataset.

[0062] Step 200: Preprocess the initial dataset by extracting ecological features, including spatial, spectral, and textural features, using a 3D convolutional neural network. Then, fuse these ecological features with topographic, geological, and meteorological auxiliary data to obtain an ecological damage feature set. Specific steps are as follows: Figure 2 As shown, it includes:

[0063] Step 201: Perform vector analysis on the initial dataset to extract key time-phase nodes of the time series;

[0064] Specifically, the difference between any two data points from different time phases in the initial dataset is performed in a multidimensional spectral space to generate a change vector. The amplitude and orientation angle of the change vector are then calculated to quantify the severity and spectral characteristics of the data changes. The amplitude and orientation angle are then filtered to identify time points where the amplitude exhibits a local peak and the orientation angle matches a preset orientation angle as key time points.

[0065] Step 202: Construct a three-dimensional convolutional neural network containing spatial convolutional layers, spectral convolutional layers, texture convolutional layers, and feature fusion layers;

[0066] Specifically, the kernel sizes of the spatial convolutional layer, the spectral convolutional layer, and the texture convolutional layer are 3×3×3, 1×1×X, and 5×5×1, respectively.

[0067] Step 203: Based on the surface deformation gradient of the mining area extracted from the key temporal nodes of the spatial convolution layer, obtain the spatial features including the boundary vector of the mining face, the expansion area of ​​the spoil heap, and the surface subsidence gradient.

[0068] Specifically, the first convolutional layer of the spatial convolutional layer uses 32 3×3×3 convolutional kernels with ReLU activation function; the second convolutional layer uses 64 3×3×3 convolutional kernels to extract the surface deformation gradient features of the mining area, expressed as: ,in For the initial dataset, G Spatial features, in this embodiment, include: the boundary vector of the mining face, the expansion area of ​​the spoil heap, and the surface settlement gradient.

[0069] Step 204: Obtain spectral features including vegetation red edge position offset, mineral feature absorption depth and water turbidity index based on the spectral derivatives of key temporal nodes extracted from the spectral convolutional layer.

[0070] Specifically, the first convolutional layer of the spectral convolutional layer uses a 1×1×8 convolutional kernel to compress the spectral dimension of the initial dataset to 1 / 4 of the original bands; the second convolutional layer uses a 1×1×3 convolutional kernel to extract the spectral derivative. The expression is: ,in and The impact data of two adjacent frames in the initial dataset. This represents the wavelength difference between two adjacent frames. The characteristic absorption depth of the mineral is then obtained from the spectral derivative. The calculation formula is: The red edge position offset, mineral characteristic absorption depth and water turbidity index are aggregated into spectral features. The water turbidity index is the ratio of the sum of wavelengths in the 450-650nm band to the sum of wavelengths in the 650-900nm band.

[0071] Step 205: Extract texture features from key temporal nodes containing surface fragmentation, fracture network density, and vegetation distribution heterogeneity index based on the texture convolutional layer;

[0072] Specifically, a primary texture of the initial dataset is first extracted using a 5×5×1 convolutional kernel to output a feature map. Then, a Gabor filter is used to enhance the multi-directional texture response of the feature map, thereby obtaining texture features that include surface fragmentation, fracture network density, and vegetation distribution heterogeneity index.

[0073] Furthermore, the high-resolution satellite imagery is converted to grayscale and divided into fixed-size grid cells. Then, the gray-level co-occurrence matrix (GLCM) is calculated within each window, and the joint probability distribution of pixel grayscale values ​​at specific directions and distances is statistically analyzed. Next, texture entropy values ​​are extracted based on the GLCM matrix. The formula for calculating the texture entropy value is: ,in Represents grayscale level i and j The co-occurrence probability is finally determined by the formula. Mapping the entropy value to the [0,1] interval yields the surface fragmentation. The closer this value is to 1, the higher the degree of surface fragmentation caused by mining activities. The fracture network density is the ratio of the total length of surface fractures to the grid area. The vegetation distribution heterogeneity index is the total area of ​​all vegetation patches within a grid cell.

[0074] Step 206: Perform feature splicing and dimensionality reduction compression on spatial features, spectral features, and texture features to obtain ecological features.

[0075] Specifically, spatial features, spectral features, and texture features are concatenated along the channel dimension to generate a fused feature map, which is then compressed to 128 dimensions using a 1×1×1 convolution kernel, and ecological features are obtained through a Sigmoid activation function and 3D convolution operation.

[0076] Step 207: Integrate ecological characteristics with topographic, geological and meteorological auxiliary data to obtain an ecological damage characteristic set.

[0077] Step 300: Based on the ecological damage feature set, construct a mine ecological damage classification model using the support vector machine algorithm, and solve the nonlinear classification problem of the mine ecological damage classification model by kernel function mapping to obtain the ecological damage type and spatial distribution;

[0078] Specifically, a historical dataset is first generated based on historical image data of the mining area and surrounding ecologically sensitive areas during the non-mining period. Then, the Gaussian radial basis function kernel is used as the kernel function of the support vector machine (SVM) model, and the SVM model parameters are optimized by maximizing the classification margin objective function to obtain an ecological damage classification model. Finally, the correlation between the ecological damage feature set and the historical dataset is calculated using the ecological damage classification model, and the ecological damage type and spatial distribution are obtained based on the correlation. In this embodiment, the ecological damage types include: vegetation destruction, soil and water pollution, land subsidence, and biodiversity loss.

[0079] Furthermore, the expression for the Gaussian radial basis kernel function is: ,in x i and x j This represents the feature vector of the dataset. The expression for the objective function is: ,in The normal vector of the classification hyperplane in the support vector machine model. b This is the bias term. The correlation between the ecological damage feature set and the historical dataset in this embodiment is represented by the Euclidean distance between the two sets. The parameter optimization process in this embodiment uses a particle swarm optimization algorithm for iterative updates.

[0080] Step 400: Calculate the ecological damage index of the mining area based on the type of ecological damage, obtain the ecological damage value by combining spatial analysis algorithm, and normalize the ecological damage value;

[0081] Specifically, based on the type of ecological damage, the ecological damage index for each spatial grid unit in the mining area is calculated using the following formula:

[0082] ;

[0083] in, These are the weighting coefficients. Let be the probability of type c damage.

[0084] The ecological damage index is then corrected using spatial autocorrelation analysis and kernel density estimation to obtain the corrected damage index. Specifically, the spatial autocorrelation index is first calculated based on the ecological damage index, using the following formula:

[0085] ;

[0086] in, This is the spatial weight matrix. n The number of neighboring grid cells. The average value of the ecological damage index is given; then, the damage diffusion density of the spatial autocorrelation index is obtained through the Gaussian kernel function, expressed as:

[0087] ;

[0088] Where h is the bandwidth, Represents a grid k With grid i The Euclidean distance.

[0089] Next, the modified damage index is obtained based on the damage diffusion density and spatial autocorrelation index. The calculation formula is as follows:

[0090] ;

[0091] in The density represents the damage diffusion density. Finally, the ecological damage value is calculated based on the modified damage index and environmental sensitivity factors, using the following formula:

[0092] ;

[0093] in, To correct the damage index, This is the adjustment coefficient for the sensitive area. Sensitive area level, For the current grid rainfall erosion force, The regional baseline erosion force was used, and the ecological damage value was normalized using the maximum-minimum normalization method.

[0094] Step 500: Classify the ecological damage level according to the normalized ecological damage value. When the ecological damage value reaches or exceeds the preset threshold, trigger a multi-level early warning mechanism and generate ecological restoration suggestions.

[0095] Specifically, when the ecological damage value is within the range of [0, 3), the ecological damage level is determined to be slight damage; when the ecological damage value is within the range of [3, 6), the ecological damage level is determined to be moderate damage; when the ecological damage value is within the range of [6, 8), the ecological damage level is determined to be severe damage; and when the ecological damage value is greater than or equal to 8, the ecological damage level is determined to be extremely severe damage. Furthermore, a dynamic threshold is generated based on environmental parameters, including rainfall erosivity index and topographic relief. The formula for calculating the dynamic threshold is:

[0096] ;

[0097] Where TI represents topographic relief and RI represents rainfall erosivity index. When the ecological damage value ≥ A blue alert is generated when the ecological damage value is ≥ max( A yellow alert is generated when the ecological damage value is ≥6 (3.5); an orange alert is generated when the ecological damage value is ≥8 (3.5); and an orange alert is generated when the ecological damage value is ≥8 (3.5). A red alert is generated when the value is greater than 0.5.

[0098] This invention also provides an intelligent identification system for mine ecological damage based on remote sensing and support vector machines, comprising:

[0099] The data acquisition module is used to simultaneously collect multi-source remote sensing data, including high-resolution satellite imagery, multispectral imagery, and radar imagery, in the mining area and surrounding ecologically sensitive areas to obtain an initial dataset.

[0100] The data preprocessing module is used to preprocess the initial dataset, extract ecological features including spatial, spectral, and textural features through a three-dimensional convolutional neural network, and fuse the ecological features with topographic, geological, and meteorological auxiliary data to obtain an ecological damage feature set.

[0101] The classification modeling module is used to construct a mine ecological damage classification model based on the ecological damage feature set using the support vector machine algorithm, and to solve the nonlinear classification problem of the mine ecological damage classification model through kernel function mapping, so as to obtain the ecological damage type and spatial distribution.

[0102] The index calculation module is used to calculate the ecological damage index of the mining area according to the type of ecological damage, and to obtain the ecological damage value by combining spatial analysis algorithm, and to normalize the ecological damage value;

[0103] The early warning generation module is used to classify the ecological damage level based on the normalized ecological damage value. When the ecological damage value reaches or exceeds the preset threshold, a multi-level early warning mechanism is triggered and ecological restoration suggestions are generated.

[0104] The beneficial effects of this invention are as follows:

[0105] 1) By simultaneously acquiring high-resolution satellite imagery, multispectral imagery, and radar imagery, and combining them with topographic / geological / meteorological auxiliary data, the problem of feature loss caused by the reliance on single remote sensing data in traditional methods has been solved, significantly improving the comprehensiveness and accuracy of ecological damage identification;

[0106] 2) The use of a three-dimensional convolutional neural network to automatically extract spatial, spectral and texture features overcomes the limitations of manually designed features and significantly improves the feature representation capability in complex scenes;

[0107] 3) A support vector machine model was introduced and combined with the Gaussian radial basis kernel function. The nonlinear classification problem of the mining environment was solved by kernel mapping, which improved the accuracy of the classification model in identifying complex damage types such as vegetation destruction and water and soil pollution.

[0108] 4) An innovative algorithm for Ecological Damage Index (EDI) and Ecological Damage Value (EDV) was designed. The spatial autocorrelation analysis introduced was used to quantify the spatial clustering of damage, and the kernel density estimation was used to characterize the damage diffusion effect, so that the assessment results are more consistent with the spatiotemporal propagation law of ecological damage in mining areas.

[0109] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0110] Specific examples have been used to illustrate the principles and implementation methods of this invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of this invention. Furthermore, those skilled in the art will recognize that, based on the ideas of this invention, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this invention.

Claims

1. A method for intelligent identification of mine ecological damage based on remote sensing and support vector machine, characterized in that, Includes the following steps: In the mining area and surrounding ecologically sensitive areas, multi-source remote sensing data including high-resolution satellite imagery, multispectral imagery and radar imagery were collected simultaneously to obtain an initial dataset. The initial dataset is preprocessed, and ecological features including spatial, spectral, and textural features are extracted using a three-dimensional convolutional neural network. These ecological features are then fused with topographic, geological, and meteorological auxiliary data to obtain an ecological damage feature set. Based on the ecological damage feature set, a mine ecological damage classification model is constructed using the support vector machine algorithm. The nonlinear classification problem of the mine ecological damage classification model is solved by kernel function mapping to obtain the ecological damage type and spatial distribution. The ecological damage index of the mining area is calculated based on the ecological damage type, and the ecological damage value is obtained by combining spatial analysis algorithm. The ecological damage value is then normalized. The ecological damage level is classified according to the normalized ecological damage value. When the ecological damage value reaches or exceeds the preset threshold, a multi-level early warning mechanism is triggered and ecological restoration suggestions are generated. The ecological damage index of the mining area is calculated based on the aforementioned ecological damage type, and the ecological damage value is obtained by combining spatial analysis algorithms. The ecological damage value is then normalized, including: Based on the aforementioned type of ecological damage, the ecological damage index is calculated for each spatial grid unit in the mining area; the formula for calculating the ecological damage index is as follows: ,in, These are the weighting coefficients. This represents the probability of type c damage. The ecological damage index was corrected by spatial autocorrelation analysis and kernel density estimation to obtain the corrected damage index; The ecological damage value is calculated based on the modified damage index and environmental sensitivity factors; the formula for calculating the ecological damage value is: ,in, To correct the damage index, This is the adjustment coefficient for the sensitive area. Sensitive area level, For the current grid rainfall erosion force, As a regional benchmark erosion force; The ecological damage value was normalized using the maximum-minimum normalization method.

2. The intelligent identification method for mine ecological damage based on remote sensing and support vector machine according to claim 1, characterized in that, In the mining area and surrounding ecologically sensitive areas, multi-source remote sensing data, including high-resolution satellite imagery, multispectral imagery, and radar imagery, were simultaneously collected to obtain an initial dataset, including: Based on the mining area and topographic diffusion model, data collection points were determined in the core mining area, transition area and ecological protection area respectively; The high-resolution satellite images of the data collection points are acquired through a satellite remote sensing platform, and land cover change information is extracted. Multispectral images of the data collection points were acquired using a drone-borne hyperspectral instrument, and vegetation stress index, water body index, and building index were calculated. Surface deformation and soil moisture were extracted from the radar images. The initial dataset is obtained by spatiotemporally aligning the land cover change information, the vegetation stress index, the water body index, the building index, the land surface deformation, and the soil moisture.

3. The intelligent identification method for mine ecological damage based on remote sensing and support vector machine according to claim 1, characterized in that, The initial dataset is preprocessed, and ecological features including spatial, spectral, and textural features are extracted using a 3D convolutional neural network. These ecological features are then fused with topographic, geological, and meteorological auxiliary data to obtain an ecological damage feature set, including: Variation vector analysis was performed on the initial dataset to extract key time-phase nodes of the time series; Construct the three-dimensional convolutional neural network comprising spatial convolutional layers, spectral convolutional layers, texture convolutional layers, and feature fusion layers; The spatial features, including the mining face boundary vector, spoil heap expansion area, and surface subsidence gradient, are obtained from the surface deformation gradient of the key temporal nodes extracted by the spatial convolution layer. The spectral features, including vegetation red edge position offset, mineral feature absorption depth, and water turbidity index, are obtained from the spectral derivatives of the key temporal nodes extracted by the spectral convolutional layer. The texture features, including surface fragmentation, fracture network density, and vegetation distribution heterogeneity index, are extracted from the key temporal nodes based on the texture convolutional layer. The ecological features are obtained by combining and compressing the spatial features, spectral features, and texture features.

4. The intelligent identification method for mine ecological damage based on remote sensing and support vector machine according to claim 3, characterized in that, Variation vector analysis is performed on the initial dataset to extract key time-phase nodes of the time series, including: The difference operation is performed on any two data points from different time phases in the initial dataset in the multidimensional spectral space to obtain the change vector; Calculate the magnitude and orientation angle of the change vector; the magnitude is used to represent the drasticness of the data change, and the orientation angle is used to represent the spectral type of the data change; Nodes where the amplitude exhibits local peaks are identified, and nodes whose direction angle is the same as a preset direction angle are selected as the key time phase nodes.

5. The intelligent identification method for mine ecological damage based on remote sensing and support vector machine according to claim 1, characterized in that, Based on the aforementioned ecological damage feature set, a mine ecological damage classification model is constructed using the support vector machine algorithm. Furthermore, the nonlinear classification problem of this model is solved through kernel function mapping, yielding the types and spatial distribution of ecological damage, including: A historical dataset is generated based on historical image data of the mining area and the surrounding ecologically sensitive area during the period before mining; The Gaussian radial basis function is used as the kernel function of the support vector machine model, and the parameters of the support vector machine model are optimized by maximizing the classification margin to obtain the ecological damage classification model. The correlation between the ecological damage feature set and the historical dataset is calculated using the ecological damage classification model, and the ecological damage type and spatial distribution are obtained based on the correlation.

6. The intelligent identification method for mine ecological damage based on remote sensing and support vector machine according to claim 1, characterized in that, The ecological damage index was corrected using spatial autocorrelation analysis and kernel density estimation to obtain a corrected damage index, including: The spatial autocorrelation index is calculated based on the ecological damage index; the formula for calculating the spatial autocorrelation index is as follows: ,in, This is the spatial weight matrix. n The number of neighboring grid cells. This represents the average value of the ecological damage index; The damage diffusion density of the spatial autocorrelation index is obtained by using a Gaussian kernel function; The corrected damage index is obtained based on the damage diffusion density and the spatial autocorrelation index; the formula for calculating the corrected damage index is: ,in This represents the damage diffusion density.

7. The intelligent identification method for mine ecological damage based on remote sensing and support vector machine according to claim 1, characterized in that, Ecological damage levels are classified based on the normalized ecological damage values. When the ecological damage value reaches or exceeds a preset threshold, a multi-level early warning mechanism is triggered and ecological restoration suggestions are generated, including: When the ecological damage value is within the range of [0,3), the ecological damage level is determined to be mild damage; when the ecological damage value is within the range of [3,6), the ecological damage level is determined to be moderate damage; when the ecological damage value is within the range of [6,8), the ecological damage level is determined to be severe damage; when the ecological damage value is greater than or equal to 8, the ecological damage level is determined to be extremely severe damage. A dynamic threshold is generated based on environmental parameters, including rainfall erosivity index and topographic relief. The formula for calculating the dynamic threshold is: Where TI is the topographic relief and RI is the rainfall erosivity index; When the ecological damage value When the ecological damage value is reached, a blue alert is generated. When the ecological damage value is ≥6, a yellow alert is generated; when the ecological damage value is ≥8, an orange alert is generated; when the ecological damage value is ≥8, an orange alert is generated. A red alert is generated at that time.

8. A smart identification system for mine ecological damage based on remote sensing and support vector machine, characterized in that, The intelligent identification method for mine ecological damage according to any one of claims 1 to 7, the system comprising: The data acquisition module is used to simultaneously collect multi-source remote sensing data, including high-resolution satellite imagery, multispectral imagery, and radar imagery, in the mining area and surrounding ecologically sensitive areas to obtain an initial dataset. The data preprocessing module is used to preprocess the initial dataset, extract ecological features including spatial features, spectral features and texture features through a three-dimensional convolutional neural network, and fuse the ecological features with topographic, geological and meteorological auxiliary data to obtain an ecological damage feature set; The classification modeling module is used to construct a mine ecological damage classification model based on the ecological damage feature set using the support vector machine algorithm, and to solve the nonlinear classification problem of the mine ecological damage classification model through kernel function mapping, so as to obtain the ecological damage type and spatial distribution. The index calculation module is used to calculate the ecological damage index of the mining area according to the ecological damage type, and obtain the ecological damage value by combining the spatial analysis algorithm, and normalize the ecological damage value. The early warning generation module is used to classify the ecological damage level according to the normalized ecological damage value, and to trigger a multi-level early warning mechanism and generate ecological restoration suggestions when the ecological damage value reaches or exceeds a preset threshold.

Images