Mine ecological environment restoration evaluation method and system based on deep learning

Abstract

The invention discloses a mine ecological environment restoration evaluation method and system based on deep learning, and the method comprises the steps: obtaining related data and evaluation indexes of a mine ecological environment restoration effect, carrying out the fuzzification of the indexes through employing a fuzzy mathematical theory, and obtaining a fuzzy evaluation index set; constructing a fuzzy evaluation index system according to the fuzzy evaluation index set, and obtaining a fuzzy score matrix of each evaluation index and fuzzy information of the fuzzy score matrix; based on fuzzy information of the fuzzy scoring matrix, high-level features of the fuzzy information are extracted; determining different evaluation grades and standards according to the clustering result, and representing the matching degree between the evaluation object and each evaluation grade by adopting a fuzzy membership degree to obtain an evaluation result; establishing a fuzzy reasoning rule base by using a fuzzy reasoning technology; and performing information fusion on the fuzzy scores of the evaluation indexes by adopting a fuzzy integral method, and comprehensively considering the interaction influence and the importance degree difference between the evaluation indexes to obtain a final comprehensive evaluation result of the mine ecological restoration effect.

Description

Technical Field

[0001] The present invention belongs to the field of information technology, and in particular relates to a method and system for evaluating mine ecological environment restoration based on deep learning. Background Art

[0002] The evaluation of the ecological environment restoration effect of mines faces the challenges of indicator ambiguity and uncertainty. Traditional evaluation methods are difficult to fully reflect the complexity of ecological restoration and are prone to subjective bias. How to build a scientific and reasonable evaluation system, accurately quantify various indicators and conduct comprehensive evaluation has become a key issue that needs to be solved urgently. Specifically, there are cross-influences between the evaluation indicators, and it is difficult to define clear weights; the evaluation criteria for different recovery stages are also different. In addition, the dynamic characteristics of the ecosystem make it difficult for static evaluation methods to capture trends in the recovery process. How to combine qualitative and quantitative analysis to achieve objective and fair evaluation under fuzzy conditions is a technical challenge faced by this field. At the same time, the interpretability and practicality of the evaluation results are also important considerations. How to transform complex mathematical models into easy-to-understand and easy-to-apply evaluation tools so that they can provide strong support for mine ecological restoration decisions is also a direction that urgently needs to be broken through. In general, establishing a set of fuzzy evaluation methods that adapt to the characteristics of mine ecological restoration and take into account both scientificity and practicality is the core issue that promotes the development of this field. Summary of the invention

[0003] In order to solve the above technical problems, the present invention proposes a mine ecological environment restoration evaluation method and system based on deep learning, which can comprehensively reflect the fuzziness and complexity characteristics of mine ecological restoration effects and improve the accuracy of evaluation.

[0004] On the one hand, to achieve the above-mentioned purpose, the present invention provides a method for evaluating the restoration of a mine ecological environment based on deep learning, comprising:

[0005] Obtain relevant data and evaluation indicators of the ecological environment restoration effect of the mine, use fuzzy mathematics theory to fuzzify the indicators, map the value range of the indicators to a preset interval, and obtain a fuzzy evaluation indicator set;

[0006] According to the fuzzy evaluation index set, a fuzzy evaluation index system is constructed, the weight of each evaluation index is determined by using the hierarchical analysis method, and the fuzzy scoring matrix of each evaluation index and the fuzzy information of the fuzzy scoring matrix are obtained by expert scoring;

[0007] Based on the fuzzy information of the fuzzy scoring matrix, a deep learning model is used to perform feature extraction and representation learning on the fuzzy information to extract high-level features of the fuzzy information;

[0008] Based on the high-level features of the fuzzy information, a fuzzy clustering algorithm is used to group the evaluation objects to determine the clustering results, different evaluation levels and standards are determined according to the clustering results, and fuzzy membership is used to characterize the matching degree between the evaluation objects and each evaluation level to obtain the evaluation results;

[0009] According to the clustering results and the evaluation results, combined with the domain expert knowledge, a fuzzy reasoning technology is used to establish a fuzzy reasoning rule base;

[0010] Based on the fuzzy reasoning rule base, the fuzzy integral method is used to fuse the fuzzy scores of the evaluation indicators, and the interactive effects and importance differences between the evaluation indicators are comprehensively considered to obtain the final comprehensive evaluation results of the mine ecological restoration effect.

[0011] On the other hand, to achieve the above-mentioned purpose, the present invention also provides a mine ecological environment restoration evaluation system based on deep learning, including: a fuzzy evaluation index acquisition module, a fuzzy scoring matrix acquisition module, a fuzzy information high-level feature extraction module, an evaluation result acquisition module, a fuzzy reasoning rule base construction module, and a final comprehensive evaluation result acquisition module;

[0012] The fuzzy evaluation index acquisition module is used to obtain relevant data and evaluation indicators of the mine ecological environment restoration effect, use fuzzy mathematics theory to fuzzify the indicators, map the value range of the indicators to a preset interval, and obtain a fuzzy evaluation index set;

[0013] The fuzzy scoring matrix acquisition module is used to construct a fuzzy evaluation index system according to the fuzzy evaluation index set, determine the weight of each evaluation index by using the hierarchical analysis method, and obtain the fuzzy scoring matrix of each evaluation index and the fuzzy information of the fuzzy scoring matrix by expert scoring;

[0014] The fuzzy information high-level feature extraction module is used to extract features and representation learning of the fuzzy information based on the fuzzy information of the fuzzy scoring matrix using a deep learning model to extract high-level features of the fuzzy information;

[0015] The evaluation result acquisition module is used to group the evaluation objects based on the high-level features of the fuzzy information using a fuzzy clustering algorithm to determine clustering results, determine different evaluation levels and standards according to the clustering results, and use fuzzy membership to characterize the matching degree between the evaluation object and each evaluation level to obtain the evaluation result;

[0016] The fuzzy reasoning rule base construction module is used to establish a fuzzy reasoning rule base using fuzzy reasoning technology based on the clustering results and the evaluation results in combination with domain expert knowledge;

[0017] The final comprehensive evaluation result acquisition module is used to fuse the fuzzy scores of each evaluation indicator based on the fuzzy reasoning rule base using a fuzzy integral method, comprehensively consider the interactive effects and importance differences between the evaluation indicators, and obtain the final comprehensive evaluation result of the mine ecological restoration effect.

[0018] Technical effect of the invention: The invention discloses a mine ecological environment restoration evaluation method and system based on deep learning, which can comprehensively reflect the fuzziness and complexity characteristics of mine ecological restoration effects and improve the accuracy of evaluation; visualize the fuzzy classification and evaluation results through heat maps or radar maps, and provide an intuitive reference for the analysis and decision-making of mine ecological restoration effects; compare the comprehensive evaluation score with the preset evaluation criteria to judge the level of mine ecological restoration effects; according to the judged mine ecological restoration effect level, put forward corresponding ecological restoration measures to guide subsequent mine ecological restoration work; through this series of data processing and analysis, the mine ecological restoration work can be evaluated and guided more scientifically. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0020] Figure 1 A schematic diagram of a process for evaluating a mine ecological environment restoration method based on deep learning according to an embodiment of the present invention;

[0021] Figure 2 This is a structural schematic diagram of a mine ecological environment restoration evaluation system based on deep learning according to an embodiment of the present invention. DETAILED DESCRIPTION

[0022] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0023] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0024] like Figure 1As shown, this embodiment provides a mine ecological environment restoration evaluation method based on deep learning, including: obtaining relevant data and evaluation indicators of the mine ecological environment restoration effect, using fuzzy mathematics theory to fuzzify the indicators, mapping the value range of the indicators to a preset interval, and obtaining a fuzzy evaluation indicator set;

[0025] According to the fuzzy evaluation index set, a fuzzy evaluation index system is constructed, the weight of each evaluation index is determined by using the hierarchical analysis method, and the fuzzy scoring matrix of each evaluation index and the fuzzy information of the fuzzy scoring matrix are obtained by expert scoring;

[0026] Based on the fuzzy information of the fuzzy scoring matrix, a deep learning model is used to perform feature extraction and representation learning on the fuzzy information to extract high-level features of the fuzzy information;

[0027] Based on the high-level features of the fuzzy information, a fuzzy clustering algorithm is used to group the evaluation objects to determine the clustering results, different evaluation levels and standards are determined according to the clustering results, and fuzzy membership is used to characterize the matching degree between the evaluation objects and each evaluation level to obtain the evaluation results;

[0028] According to the clustering results and the evaluation results, combined with the domain expert knowledge, a fuzzy reasoning technology is used to establish a fuzzy reasoning rule base;

[0029] Based on the fuzzy reasoning rule base, the fuzzy integral method is used to fuse the fuzzy scores of the evaluation indicators, and the interactive effects and importance differences between the evaluation indicators are comprehensively considered to obtain the final comprehensive evaluation results of the mine ecological restoration effect.

[0030] Furthermore, in view of the fuzziness and uncertainty of the evaluation indicators in the fuzzy evaluation indicator set, the fuzzy evaluation results of the mine ecological environment restoration effect are obtained, including:

[0031] Aiming at the fuzziness and uncertainty of the evaluation indicators in the fuzzy evaluation indicator set, fuzzy mathematical theory is adopted to establish a fuzzy evaluation indicator model, and the value range of the evaluation indicator is mapped to a preset interval to complete the fuzzy representation of the evaluation indicator;

[0032] According to the fuzzy evaluation index model, a fuzzy comprehensive evaluation method is adopted to calculate the membership degree of each evaluation index to obtain the fuzzy evaluation value of the evaluation index;

[0033] Obtaining a target value of an environmental restoration effect, and performing fuzzification processing on the target value to obtain a fuzzy target value;

[0034] The degree between the fuzzy evaluation value and the fuzzy target value is calculated, and the fuzzy similarity method is used to obtain the fuzzy evaluation result of the mine ecological environment restoration effect.

[0035] Furthermore, the fuzzy scoring matrix of each evaluation index and the fuzzy information of the fuzzy scoring matrix are obtained, including:

[0036] The weight of each evaluation index was determined through the hierarchical analysis method; assuming that the evaluation indicators of the ecological environment restoration effect of the mine include vegetation coverage, soil heavy metal content and water quality index, through expert consultation and data analysis, these indicators were given weights of 5, 3 and 2 respectively.

[0037] Through expert scoring, the fuzzy scoring matrix data of each indicator is obtained; the fuzzy scoring of vegetation coverage may be {7 / good, 2 / medium, 1 / poor}, the fuzzy scoring of soil heavy metal content may be {6 / qualified, 4 / exceeding the standard}, and the fuzzy scoring of water quality index may be {8 / excellent, 1 / good, 1 / poor}.

[0038] Comprehensively consider the impact of various uncertain factors on the score and process the fuzzy score matrix data;

[0039] The fuzzy comprehensive evaluation method is used to perform weighted calculation on the indicator weight value and the processed fuzzy scoring matrix to obtain the comprehensive evaluation result; the comprehensive evaluation result may be {75 / good, 15 / medium, 1 / poor}.

[0040] According to the preset threshold 7, the comprehensive evaluation results are classified to obtain the fuzzy scoring matrix of each evaluation index and the fuzzy information of the fuzzy scoring matrix; if the score of soil heavy metal content frequently exceeds the standard, cluster analysis may point out that this index is the focus of improvement. According to the results of cluster analysis, targeted improvement measures are proposed, such as increasing the use of soil conditioners, adjusting vegetation planting structure, and dynamically adjusting the fuzzy evaluation index system and weights to optimize the fuzzy evaluation process and achieve continuous improvement of the mine ecological environment.

[0041] Furthermore, extracting high-level features of the fuzzy information includes:

[0042] Obtain fuzzy scoring matrix data, preprocess the fuzzy scoring information in the matrix, and convert the scoring data into a format suitable for deep learning model input;

[0043] A convolutional neural network is used to extract features from the preprocessed fuzzy score data, and the initial high-level feature representation is obtained through multi-layer nonlinear transformation of convolutional layers and pooling layers;

[0044] The initial high-level features extracted by the convolutional neural network are input into the long short-term memory network, which captures the long-distance dependencies and temporal features in the fuzzy rating data through the gating mechanism and memory units.

[0045] In the output layer of the LSTM network, the attention mechanism is used to perform weighted aggregation on the features of different time steps to obtain the overall feature representation of the fuzzy rating data;

[0046] The overall feature representation of the fuzzy rating data is input into a fully connected neural network, and the fuzzy relationship is modeled and mined through multi-layer nonlinear transformation to obtain the correlation and influence relationship between different rating objects;

[0047] According to the mined fuzzy relations, the fuzzy reasoning method is used to predict and evaluate the new scoring objects, and the position and correlation of the new scoring objects in the fuzzy scoring matrix are obtained;

[0048] The result of fuzzy reasoning is integrated with the original fuzzy scoring matrix, the fuzzy scoring matrix is ​​updated and optimized, and high-level features of the fuzzy information are extracted.

[0049] Specifically, first, the fuzzy rating matrix data is preprocessed, such as using the Z-score standardization method to standardize the rating data to ensure the consistency and comparability of the data before entering the deep learning model. Then, the convolutional neural network (CNN) is used to extract features from the preprocessed data. By setting multiple convolutional layers and pooling layers, three convolutional layers are used, each containing 32 3x3 convolution kernels, and the corresponding maximum pooling layer to automatically learn the high-level feature representation in the data. Then, the features extracted by CNN are input into the long short-term memory network (LSTM), and the number of LSTM units is set to 64. Through its gating mechanism and memory units, the temporal features and long-distance dependencies in the rating data are effectively captured. In the output layer of the LSTM network, the attention mechanism is introduced to obtain the overall feature representation of the rating data by calculating the weighted sum of the outputs at different time steps, and the weights are automatically learned by data-driven learning. This overall feature representation is then input into a fully connected neural network consisting of three layers, each containing 128, 64 and 32 neurons, respectively, and nonlinear transformation is performed through the ReLU activation function to further mine the fuzzy relationship. Finally, based on the mined fuzzy relationships, fuzzy reasoning methods, such as fuzzy logic controllers, are used to predict and evaluate new scoring objects. The fuzzy scoring matrix is ​​updated according to the reasoning results to improve the accuracy and interpretability of the scoring, thereby providing more precise support for decision analysis.

[0050] Furthermore, the evaluation results include:

[0051] According to the obtained high-level features, the fuzzy C-means clustering algorithm is used to group and cluster the evaluation objects to obtain different clusters;

[0052] According to the characteristic distribution of the evaluation objects within each cluster, the corresponding evaluation level and evaluation standard are determined;

[0053] Calculate the Euclidean distance between each evaluation object and the center of each cluster, and convert the distance into a membership value through the fuzzy membership function;

[0054] According to the membership value of the assessment object, determine the matching degree between the membership value and each assessment level, and classify the membership value into the assessment level with the highest matching degree;

[0055] Taking into account the distribution of membership of the assessment object at each assessment level, the fuzzy classification and comprehensive evaluation of the mine ecological restoration effect are carried out to obtain the evaluation results.

[0056] Specifically, the input data is 1000 samples, each sample contains 10 features. After CNN processing, a 128-dimensional high-level feature vector is extracted. The extracted high-level features are grouped and clustered using the fuzzy C-means (FCM) clustering algorithm. The number of clusters is set to 4, and the initial cluster center is randomly generated. The cluster center and the membership matrix are updated through iterative calculation until the convergence condition is met, and finally 4 clusters are obtained. Each cluster represents a different level of ecological restoration effect, such as "excellent", "good", "medium" and "poor". For each cluster, the characteristic distribution of the internal evaluation object is analyzed to determine the corresponding evaluation level and evaluation criteria. For example, the mean vegetation coverage rate in cluster 1 reaches 85%, and the mean soil quality index exceeds 8. It is defined as the "excellent" level, and the corresponding evaluation standard threshold is set. For each evaluation object, the Euclidean distance from the center of each cluster is calculated. Assume that the high-level feature vector of an evaluation object is \[7,9,6,\ldots,8\], and calculate its distance from the center of the cluster, which is 1, 5, 0, and 2 respectively. The distance is converted into a membership value through the fuzzy membership function. Using the Gaussian membership function, the membership of the object to the four clusters is obtained to be 3, 7, 1, and 05 respectively. According to the membership value of the evaluation object, its matching degree with each evaluation level is judged, and it is divided into the evaluation level with the highest matching degree. For example, the object has the highest membership of the "good" level (7), so it is divided into the "good" level. Considering the membership distribution of the evaluation object at each evaluation level, the ecological restoration effect of the mine is fuzzily classified and comprehensively evaluated. A comprehensive evaluation model is constructed, and the weights are assigned to the vegetation coverage rate 4, the soil quality index 3, and the water quality index 3. The comprehensive membership is calculated, and the comprehensive evaluation of the object is finally determined to be "good". The fuzzy classification and evaluation results are visualized, and the membership distribution of each evaluation object at different levels is displayed using heat maps and radar maps. The depth of color represents the degree of membership, which directly reflects the distribution of the ecological restoration effect of the mine. The darker color of a certain area in the heat map indicates that the ecological restoration effect of the area is better, providing an intuitive reference for decision makers. Through the above steps, the automated evaluation and visualization of the ecological restoration effect of the mine are realized, providing a scientific basis for the subsequent formulation of ecological restoration strategies and resource allocation.

[0057] Furthermore, establishing a fuzzy reasoning rule base includes:

[0058] According to the fuzzy classification and evaluation results, domain expert knowledge is obtained and an expert knowledge base is established;

[0059] Using fuzzy reasoning technology, based on the expert knowledge base, a fuzzy evaluation rule base is constructed; from the fuzzy evaluation rule base, a series of IF-THEN form of fuzzy reasoning rules are extracted;

[0060] Obtain relevant data and information for the mine ecological restoration cases to be evaluated;

[0061] Input case data and information into fuzzy reasoning rules, and obtain fuzzy evaluation results of mine ecological restoration effects through fuzzy reasoning process;

[0062] According to the fuzzy evaluation results, the weighted average algorithm is used to calculate the comprehensive evaluation score of the mine ecological restoration effect;

[0063] Based on the comprehensive evaluation score and the preset decision threshold, the level of mine ecological restoration effect is judged and a fuzzy reasoning rule library is established.

[0064] Furthermore, the final comprehensive evaluation results of the mine ecological restoration effect include:

[0065] According to the established fuzzy reasoning rule base, the fuzzy scores of the evaluation indicators are obtained;

[0066] The fuzzy integral method is used to carry out information fusion processing on the fuzzy scores of evaluation indicators;

[0067] When information is fused, it is determined whether there is any interaction between the indicators. If there is any interaction, the influence is comprehensively considered in the fusion process;

[0068] When information is fused, the importance differences of each evaluation index are obtained, and a higher weight coefficient is assigned to the index with higher importance;

[0069] The final comprehensive evaluation results of mine ecological restoration effects were obtained through the fuzzy integral method.

[0070] like Figure 2 As shown, this embodiment also provides a mine ecological environment restoration evaluation system based on deep learning, including: a fuzzy evaluation index acquisition module, a fuzzy scoring matrix acquisition module, a fuzzy information high-level feature extraction module, an evaluation result acquisition module, a fuzzy reasoning rule base construction module, and a final comprehensive evaluation result acquisition module;

[0071] The fuzzy evaluation index acquisition module is used to obtain relevant data and evaluation indicators of the mine ecological environment restoration effect, use fuzzy mathematics theory to fuzzify the indicators, map the value range of the indicators to the preset interval, and obtain the fuzzy evaluation indicator set;

[0072] The fuzzy scoring matrix acquisition module is used to construct a fuzzy evaluation index system based on the fuzzy evaluation index set, determine the weight of each evaluation index using the hierarchical analysis method, and obtain the fuzzy scoring matrix of each evaluation index and the fuzzy information of the fuzzy scoring matrix through expert scoring;

[0073] The fuzzy information high-level feature extraction module is used for fuzzy information based on the fuzzy scoring matrix. It uses a deep learning model to perform feature extraction and representation learning on the fuzzy information to extract high-level features of the fuzzy information.

[0074] The evaluation result acquisition module is used to group the evaluation objects based on the high-level features of fuzzy information and use the fuzzy clustering algorithm to determine the clustering results, determine different evaluation levels and standards according to the clustering results, and use the fuzzy membership degree to represent the matching degree between the evaluation object and each evaluation level to obtain the evaluation result;

[0075] The fuzzy reasoning rule base construction module is used to establish a fuzzy reasoning rule base based on clustering results and evaluation results, combined with domain expert knowledge, using fuzzy reasoning technology;

[0076] The final comprehensive evaluation result acquisition module is used to fuse the fuzzy scores of each evaluation index based on the fuzzy reasoning rule library and the fuzzy integral method, comprehensively consider the interaction and importance differences between the evaluation indexes, and obtain the final comprehensive evaluation results of the mine ecological restoration effect.

[0077] The present invention discloses a mine ecological environment restoration evaluation method and system based on deep learning, which can comprehensively reflect the fuzziness and complexity characteristics of mine ecological restoration effects and improve the accuracy of evaluation; visualize the fuzzy classification and evaluation results through heat maps or radar maps to provide intuitive reference for analysis and decision-making of mine ecological restoration effects; compare the comprehensive evaluation score with the preset evaluation criteria to judge the level of mine ecological restoration effects; and propose corresponding ecological restoration measures based on the judged mine ecological restoration effect level to guide subsequent mine ecological restoration work; through this series of data processing and analysis, the mine ecological restoration work can be evaluated and guided more scientifically.

[0078] The above are only preferred specific implementations of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A mine ecological environment restoration evaluation method based on deep learning, characterized in that: include: Obtain relevant data and evaluation indicators of the ecological environment restoration effect of the mine, use fuzzy mathematics theory to fuzzify the indicators, map the value range of the indicators to a preset interval, and obtain a fuzzy evaluation indicator set; According to the fuzzy evaluation index set, a fuzzy evaluation index system is constructed, the weight of each evaluation index is determined by using the hierarchical analysis method, and the fuzzy scoring matrix of each evaluation index and the fuzzy information of the fuzzy scoring matrix are obtained by expert scoring; Based on the fuzzy information of the fuzzy scoring matrix, a deep learning model is used to perform feature extraction and representation learning on the fuzzy information to extract high-level features of the fuzzy information; Based on the high-level features of the fuzzy information, a fuzzy clustering algorithm is used to group the evaluation objects to determine the clustering results, different evaluation levels and standards are determined according to the clustering results, and fuzzy membership is used to characterize the matching degree between the evaluation objects and each evaluation level to obtain the evaluation results; According to the clustering results and the evaluation results, combined with the domain expert knowledge, a fuzzy reasoning technology is used to establish a fuzzy reasoning rule base; Based on the fuzzy reasoning rule base, the fuzzy integral method is used to fuse the fuzzy scores of the evaluation indicators, and the interactive effects and importance differences between the evaluation indicators are comprehensively considered to obtain the final comprehensive evaluation results of the mine ecological restoration effect.

2. The deep learning-based mine ecological environment restoration evaluation method according to claim 1, characterized in that: In view of the fuzziness and uncertainty of the evaluation indicators in the fuzzy evaluation indicator set, the fuzzy evaluation results of the mine ecological environment restoration effect are obtained, including: Aiming at the fuzziness and uncertainty of the evaluation indicators in the fuzzy evaluation indicator set, fuzzy mathematical theory is adopted to establish a fuzzy evaluation indicator model, and the value range of the evaluation indicator is mapped to a preset interval to complete the fuzzy representation of the evaluation indicator; According to the fuzzy evaluation index model, a fuzzy comprehensive evaluation method is adopted to calculate the membership degree of each evaluation index to obtain the fuzzy evaluation value of the evaluation index; Obtaining a target value of an environmental restoration effect, and performing fuzzification processing on the target value to obtain a fuzzy target value; The degree between the fuzzy evaluation value and the fuzzy target value is calculated, and the fuzzy similarity method is used to obtain the fuzzy evaluation result of the mine ecological environment restoration effect.

3. The method for evaluating the restoration of the mine ecological environment based on deep learning according to claim 1, characterized in that: Obtaining the fuzzy scoring matrix of each evaluation index and the fuzzy information of the fuzzy scoring matrix includes: The weight of each evaluation index is determined by analytic hierarchy process; Obtain the fuzzy scoring matrix data of each indicator through expert scoring; Comprehensively consider the impact of various uncertain factors on the score and process the fuzzy score matrix data; The fuzzy comprehensive evaluation method is used to perform weighted calculation on the index weight value and the processed fuzzy scoring matrix to obtain the comprehensive evaluation result; According to the preset threshold, the comprehensive evaluation results are classified to obtain the fuzzy scoring matrix of each evaluation index and the fuzzy information of the fuzzy scoring matrix.

4. The deep learning-based mine ecological environment restoration evaluation method according to claim 1, characterized in that: The high-level features for extracting the fuzzy information include: Obtain fuzzy scoring matrix data, preprocess the fuzzy scoring information in the matrix, and convert the scoring data into a format suitable for deep learning model input; A convolutional neural network is used to extract features from the preprocessed fuzzy score data, and the initial high-level feature representation is obtained through multi-layer nonlinear transformation of convolutional layers and pooling layers; The initial high-level features extracted by the convolutional neural network are input into the long short-term memory network, which captures the long-distance dependencies and temporal features in the fuzzy rating data through the gating mechanism and memory units. In the output layer of the LSTM network, the attention mechanism is used to perform weighted aggregation on the features of different time steps to obtain the overall feature representation of the fuzzy rating data; The overall feature representation of the fuzzy rating data is input into a fully connected neural network, and the fuzzy relationship is modeled and mined through multi-layer nonlinear transformation to obtain the correlation and influence relationship between different rating objects; According to the mined fuzzy relations, the fuzzy reasoning method is used to predict and evaluate the new scoring objects, and the position and correlation of the new scoring objects in the fuzzy scoring matrix are obtained; The result of fuzzy reasoning is integrated with the original fuzzy scoring matrix, the fuzzy scoring matrix is ​​updated and optimized, and high-level features of the fuzzy information are extracted.

5. The method for evaluating the restoration of the mine ecological environment based on deep learning according to claim 1, characterized in that: The evaluation results include: According to the obtained high-level features, the fuzzy C-means clustering algorithm is used to group and cluster the evaluation objects to obtain different clusters; According to the characteristic distribution of the evaluation objects within each cluster, the corresponding evaluation level and evaluation standard are determined; Calculate the Euclidean distance between each evaluation object and the center of each cluster, and convert the distance into a membership value through the fuzzy membership function; According to the membership value of the assessment object, determine the matching degree between the membership value and each assessment level, and classify the membership value into the assessment level with the highest matching degree; Taking into account the distribution of membership of the assessment object at each assessment level, the fuzzy classification and comprehensive evaluation of the mine ecological restoration effect are carried out to obtain the evaluation results.

6. The method for evaluating the restoration of the mine ecological environment based on deep learning according to claim 1, characterized in that: Building a fuzzy reasoning rule base includes: According to the fuzzy classification and evaluation results, domain expert knowledge is obtained and an expert knowledge base is established; Using fuzzy reasoning technology, based on the expert knowledge base, a fuzzy evaluation rule base is constructed; from the fuzzy evaluation rule base, a series of IF-THEN form of fuzzy reasoning rules are extracted; Obtain relevant data and information for the mine ecological restoration cases to be evaluated; Input case data and information into fuzzy reasoning rules, and obtain fuzzy evaluation results of mine ecological restoration effects through fuzzy reasoning process; According to the fuzzy evaluation results, the weighted average algorithm is used to calculate the comprehensive evaluation score of the mine ecological restoration effect; Based on the comprehensive evaluation score and the preset decision threshold, the level of mine ecological restoration effect is judged and a fuzzy reasoning rule library is established.

7. The deep learning-based mine ecological environment restoration evaluation method according to claim 1, characterized in that: The final comprehensive evaluation results of the mine ecological restoration effect include: According to the established fuzzy reasoning rule base, the fuzzy scores of the evaluation indicators are obtained; The fuzzy integral method is used to carry out information fusion processing on the fuzzy scores of evaluation indicators; When information is fused, it is determined whether there is any interaction between the indicators. If there is any interaction, the influence is comprehensively considered in the fusion process; When information is fused, the importance differences of each evaluation index are obtained, and a higher weight coefficient is assigned to the index with higher importance; The final comprehensive evaluation results of mine ecological restoration effects were obtained through the fuzzy integral method.

8. A system for evaluating the mine ecological environment restoration method based on deep learning according to any one of claims 1 to 7, characterized in that: include: Fuzzy evaluation index acquisition module, fuzzy scoring matrix acquisition module, fuzzy information high-level feature extraction module, evaluation result acquisition module, fuzzy reasoning rule base construction module, and final comprehensive evaluation result acquisition module; The fuzzy evaluation index acquisition module is used to obtain relevant data and evaluation indicators of the mine ecological environment restoration effect, use fuzzy mathematics theory to fuzzify the indicators, map the value range of the indicators to a preset interval, and obtain a fuzzy evaluation index set; The fuzzy scoring matrix acquisition module is used to construct a fuzzy evaluation index system according to the fuzzy evaluation index set, determine the weight of each evaluation index by using the hierarchical analysis method, and obtain the fuzzy scoring matrix of each evaluation index and the fuzzy information of the fuzzy scoring matrix by expert scoring; The fuzzy information high-level feature extraction module is used to extract features and representation learning of the fuzzy information based on the fuzzy information of the fuzzy scoring matrix using a deep learning model to extract high-level features of the fuzzy information; The evaluation result acquisition module is used to group the evaluation objects based on the high-level features of the fuzzy information using a fuzzy clustering algorithm to determine clustering results, determine different evaluation levels and standards according to the clustering results, and use fuzzy membership to characterize the matching degree between the evaluation object and each evaluation level to obtain the evaluation result; The fuzzy reasoning rule base construction module is used to establish a fuzzy reasoning rule base using fuzzy reasoning technology based on the clustering results and the evaluation results in combination with domain expert knowledge; The final comprehensive evaluation result acquisition module is used to fuse the fuzzy scores of each evaluation indicator based on the fuzzy reasoning rule base using a fuzzy integral method, comprehensively consider the interactive effects and importance differences between the evaluation indicators, and obtain the final comprehensive evaluation result of the mine ecological restoration effect.

Images