Mine geological environment evaluation system based on multi-scale characteristics and evaluation method thereof

Abstract

The invention discloses a mine geological environment evaluation system based on multi-scale characteristics and an evaluation method thereof. The mine geological environment evaluation system comprises a remote sensing image acquisition unit, a mine geological environment problem analysis unit, a mine geological environment element extraction unit, a mine geological environment problem identification unit and a mine geological environment evaluation unit. The method comprises the following steps: collecting a remote sensing image of a target mine area, extracting mine geological environment elements, analyzing temporal and spatial change characteristics of the elements, performing similarity comparison calculation with change characterization of different mine geological environment problems, identifying the geological environment problems, and training an evaluation model to obtain a geological environment evaluation prediction result. According to the method, index factor characterization of different mine geological environment problems is analyzed from remote sensing image modal data, through automatic comparison and matching with target remote sensing image multi-scale features, the problem that the efficiency and quality of a mine geological environment investigation mode and a sample selection method are low is solved, and scientific evaluation of the mine geological environment is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of mine geological environment assessment, and more specifically, to a mine geological environment assessment system and an assessment method thereof based on multi-scale features. Background Art

[0002] With the rapid development of the economy and society, the development and utilization of mining resources has become an important basic guarantee for economic development and social progress. At the same time, the development and utilization of mining resources has inevitably brought about changes and damage to the environment. The development and utilization of mineral resources has brought about environmental problems such as soil erosion, soil pollution and collapse, and has also posed a threat to human survival and development in the area affected by the mines, seriously affecting the development and utilization of mines and the sustainable development of the regional economy. It is urgent to manage and repair the ecological environment of mines. Among them, the evaluation of the quality of the geological environment of mines is an important evaluation method for the impact of the development of mining resources, and it is also the basis for subsequent management and restoration, and it is an important part of the sustainable development and utilization of mines. Therefore, the evaluation of the quality of the geological environment of mines is necessary.

[0003] At present, the commonly used evaluation methods include grid method, vector polygon method, fuzzy comprehensive evaluation method, etc., among which there are relatively more disadvantages between different methods. For example, the evaluation results of the grid method are poorly consistent with the actual boundaries. Although it is good for predicting development trends and can reflect more information, it is greatly affected by human experience and is not easy to operate, making the evaluation results difficult to be accepted by others. The evaluation results of the vector polygon method are highly consistent with the actual boundaries, can reflect more information, are less affected by human experience, and are easier to operate, but the prediction of development trends is slightly worse. The weights and membership of the fuzzy comprehensive evaluation method are difficult to determine. Therefore, it is necessary to analyze the causes and influencing factors of various environmental geological problems, and how to use the multi-scale features of the analysis based on machine learning methods for mining geological environment evaluation and prediction is an urgent problem to be solved. Summary of the invention

[0004] In order to solve the above technical problems, the present invention proposes a mine geological environment evaluation system and an evaluation method based on multi-scale features, aiming to solve the problems of low efficiency and quality of manual training sample selection methods, realize scientific evaluation of mine geological environment, and provide decision support for mine restoration and management work.

[0005] The present invention provides a mine geological environment evaluation system based on multi-scale features, comprising: a remote sensing image acquisition unit, a mine geological environment problem analysis unit, a mine geological environment element extraction unit, a mine geological environment problem identification unit and a mine geological environment evaluation unit; The remote sensing image acquisition unit is responsible for collecting remote sensing images of the target mining area, preprocessing the remote sensing images, and generating a measurement area change sequence of the target mining area; The mining geological environment problem analysis unit is responsible for retrieving mining geological environment problem examples, extracting index factors according to their cause mechanisms, and obtaining the change representation of index factors of different mining geological environment problems under remote sensing image modality; The mining geological environment element extraction unit is responsible for extracting multi-scale features in the measurement area change sequence for semantic segmentation, obtaining mining geological environment elements, and analyzing the temporal and spatial change characteristics of the mining geological environment elements; The mine geological environment problem identification unit is responsible for performing similarity comparison calculation between the temporal and spatial variation characteristics of mine geological environment elements and the variation representations of different mine geological environment problems, and identifying the geological environment problems in the target mine area; The mining geological environment evaluation unit is responsible for training the evaluation model based on the extreme learning machine, reading the index parameters of the geological environment problems in the target mining area as model input, and obtaining the evaluation prediction results of the geological environment of the target mining area.

[0006] In this solution, the remote sensing image acquisition unit is specifically: The remote sensing image acquisition unit includes an acquisition unit and a processing unit. The acquisition unit includes a remote sensing imaging device arranged in a satellite. The remote sensing imaging device is used to collect remote sensing image data of a target mining area within a preset time step from a preset remote sensing data source satellite. The processing unit performs image selection, orthophoto processing and image fusion in the remote sensing image data, obtains a high-resolution color image, obtains an aerial photograph of the target mining area as a reference image, crops the high-resolution color image based on the reference image, obtains a satellite image that meets the scope of the mission area, and after completing the remote sensing image preprocessing, generates a remote sensing image sequence according to the acquisition timestamp; In the remote sensing image sequence, the changes of remote sensing images in different periods are analyzed, the change area is marked, the change spots are extracted in the change area, the change spots are interpreted, label information is generated, and a change sequence of spots with label information is obtained; The patch change sequence is aggregated to generate a measurement area change sequence of the target mining area.

[0007] In this scheme, the mining geological environment problem analysis unit is specifically: Obtaining examples of mining geological environment problems through big data retrieval, segmenting the examples of mining geological environment problems, extracting corresponding expression form keywords and contexts containing expression form keywords, and using semantic analysis to obtain the expression form characteristics of each mining geological environment problem; Clustering the mining geological environment problem instances according to the manifestation characteristics, setting mining geological environment problem type labels for the clustering clusters corresponding to the clustering results, and reading the causal mechanism of each instance sample in each cluster; Integrate the genetic mechanisms, calculate the correlation between different genetic mechanisms and the correlation between the genetic mechanisms and the types of mining geological environmental problems, and select indicator factors according to two different correlations; The remote sensing image modal data in the example of mining geological environment problems are extracted, and the indicator factors are used to locate them in the remote sensing image modal data to obtain the change representation of the indicator factors of different mining geological environment problems under the remote sensing image modality.

[0008] In this solution, the mining geological environment element extraction unit is specifically: The mining geological environment element extraction unit includes a mining geological environment element extraction model, which is constructed based on an improved U-Net and uses Res2Net50 to replace the original encoder module; The survey area change sequence of the target mining area is used as the model input. In the encoder module Res2Net50, hierarchical residual connections are used to obtain feature maps of different scales of the change spots in the survey area change sequence. The encoded feature maps of different scales are further downsampled and propagated, and residual jump connections are used as part of the input of the decoder module. In the encoder module, the input feature map is restored through upsampling operation, and the restored feature map of each layer is used as local information to generate prediction branches. Each prediction branch is fused to achieve the integration of global information. The integrated global information is processed by parallel channel attention modules and spatial attention modules to obtain different channel descriptions. The two channel descriptions are added together to achieve fusion, and the fused feature map is identified and extracted from the mining geological environment elements through the Sigmoid layer.

[0009] In this solution, in the mining geological environment element extraction unit, the temporal and spatial variation characteristics of the mining geological environment elements are analyzed, specifically: According to the identification and extraction results of the mining geological environment elements in the measurement area change sequence, the marked mining geological environment elements are divided into grids, the divided grids are assigned values ​​through the measurement area change sequence, and the time series correlation of the grid values ​​is obtained using a time convolutional network; Constructing time edges between grids corresponding to mining geological environment elements according to the temporal correlation, and constructing space edges according to the distances between grids, and constructing a space-time graph for the grids of mining geological environment elements using the time edges and space edges; The spatiotemporal graph is represented in the form of an adjacency matrix, the adjacency matrix is ​​learned through a graph neural network, and the features are updated using information transmission and neighbor aggregation mechanisms to obtain the spatiotemporal change characteristics of the mining geological environment elements in the survey area change sequence.

[0010] In this solution, the mine geological environment problem identification unit is specifically: The dynamic time warping algorithm is used to calculate the spatiotemporal variation characteristics of each mining geological environment element and the variation representation of different index factors under remote sensing image modality, and the DTW distance between the feature change sequences is generated. The remote sensing image change representations of the indicator factors with the minimum DTW distance of each mining geological environment element are compared, the selected remote sensing image change representations of the indicator factors are aggregated, and the mining geological environment problems that meet the preset similarity standards are obtained by similarity calculation.

[0011] In this scheme, the mining geological environment assessment unit is specifically: Extracting index parameters from examples of mining geological environment problems based on the selected index factors, quantifying the index parameters, obtaining weight information of different index factors through the hierarchical analysis method, and performing hierarchical weighted calculations on the index factors using the weight information to obtain evaluation results of mining geological environment problems; An evaluation model is constructed based on the extreme learning machine, and the mining geological environment problems and their evaluation results are used as the data source of the evaluation model for training and testing to obtain the accuracy data of the evaluation model. When the accuracy meets the preset accuracy requirements, the evaluation model is output; The index factors corresponding to the mine geological environment problems in the target mine area are extracted, and the index parameters of the target mine area are read according to the survey area change sequence of the target mine area based on the index factors, and are imported into the evaluation model as model input to output the evaluation prediction results of the target mine area.

[0012] A second aspect of the present invention provides a method for evaluating a mine geological environment based on multi-scale features, comprising the following steps: Acquire a remote sensing image of a target mining area as a data source, pre-process the remote sensing image, and generate a measurement area change sequence of the target mining area; Retrieve and obtain mining geological environmental problems, extract indicator factors according to the causal mechanism of the mining geological environmental problems, and obtain the change representation of the indicator factors of different mining geological environmental problems under the remote sensing image mode; Extracting multi-scale features from the measurement area change sequence for semantic segmentation, obtaining mining geological environment elements, and analyzing the spatiotemporal change characteristics of the mining geological environment elements; Compare and calculate the similarity between the temporal and spatial variation characteristics of mining geological environmental elements and the variation representations of different mining geological environmental problems to identify the geological environmental problems in the target mining area; Based on the extreme learning machine, the evaluation model is trained using the indicator factors of different mining geological environmental problems. The indicator parameters of the geological environmental problems in the target mining area are read as model input to obtain the evaluation and prediction results of the geological environment of the target mining area.

[0013] Compared with the prior art, the present invention has the following beneficial effects: The present invention analyzes the indicator factor representation of different mine geological environment problems based on remote sensing image modal data, solves the problems of low efficiency and quality of mine geological environment survey methods and sample selection methods through automatic comparison and matching of multi-scale features of remote sensing images, and displays the spatial distribution differences of mine geological environment evolution in different periods according to the temporal evolution law of mine geological environment in different periods in combination with spatial relationships, reads temporal and spatial characteristics, identifies mine geological environment problems, realizes scientific evaluation of mine geological environment based on indicator factors of mine geological environment problems, improves evaluation accuracy, and provides a basis for mine management and restoration. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the technical solutions in the embodiments or exemplary embodiments of the present invention, the drawings required for use in the embodiments or exemplary descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained according to the drawings without paying creative work.

[0015] Figure 1 The architecture diagram of the mine geological environment evaluation system based on multi-scale features is shown; Figure 2 A flowchart of extracting mining geological environment elements in a target mining area is shown in an embodiment; Figure 3 A flowchart of obtaining a mine geological environment evaluation result in an embodiment is shown; Figure 4 A flow chart of the mine geological environment evaluation method based on multi-scale features is shown. DETAILED DESCRIPTION

[0016] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0017] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.

[0018] Figure 1 The architecture diagram of the mine geological environment evaluation system based on multi-scale features is shown; like Figure 1As shown, the embodiment of the present invention provides a mine geological environment evaluation system 1 based on multi-scale features, including: a remote sensing image acquisition unit 102, a mine geological environment problem analysis unit 104, a mine geological environment element extraction unit 106, a mine geological environment problem identification unit 108 and a mine geological environment evaluation unit 110; The remote sensing image acquisition unit 102 is responsible for collecting remote sensing images of the target mining area, preprocessing the remote sensing images, and generating a measurement area change sequence of the target mining area; The mining geological environment problem analysis unit 104 is responsible for retrieving mining geological environment problem instances, extracting index factors according to their cause mechanisms, and obtaining the change representation of index factors of different mining geological environment problems under remote sensing image modalities; The mining geological environment element extraction unit 106 is responsible for extracting multi-scale features in the measurement area change sequence for semantic segmentation, obtaining mining geological environment elements, and analyzing the spatiotemporal change characteristics of the mining geological environment elements; The mine geological environment problem identification unit 108 is responsible for performing similarity comparison calculation between the temporal and spatial variation characteristics of the mine geological environment elements and the variation representations of different mine geological environment problems, and identifying the geological environment problems in the target mine area; The mine geological environment evaluation unit 110 is responsible for training an evaluation model based on an extreme learning machine, reading index parameters of geological environmental problems in a target mine area as model input, and obtaining evaluation prediction results of the geological environment in the target mine area.

[0019] It should be noted that the remote sensing image acquisition unit includes an acquisition unit and a processing unit. The acquisition unit includes a remote sensing imaging device set in a satellite, and the remote sensing imaging device is used to collect remote sensing image data of the target mining area within a preset time step from a preset remote sensing data source satellite; the processing unit performs image selection, orthophoto processing and image fusion in the remote sensing image data, wherein a rational function regional network adjustment model is used to complete high-precision geometric processing of remote sensing images, control points are densely and evenly distributed in each image, and the error in the plain is controlled within 1 pixel, and the error in the mountainous area is controlled within 2 pixels. The Pansharp method is used to perform fusion operations on multispectral images and panchromatic images to obtain high-resolution color images; in addition, the original images are appropriately enhanced to improve image contrast and clarity. Aerial photographs of the target mining area are obtained as reference images, and the aerial photographs meet the absolute positioning accuracy of better than 10m and the accuracy requirement of 0.8m. Based on the reference images, high-resolution color images are cropped to obtain satellite images that meet the scope of the mission area. After completing the remote sensing image preprocessing, a remote sensing image sequence is generated according to the acquisition timestamp. Changes of remote sensing images in different periods are analyzed in the remote sensing image sequence, and the changed areas are marked. Change spots are extracted in the changed areas, and the changed spots are interpreted to generate label information to obtain a spot change sequence with label information. The spot change sequence is aggregated to generate a measurement area change sequence of the target mining area.

[0020] It should be noted that, in the mining geological environment problem analysis unit, mining geological environment problem instances are obtained through big data retrieval, the description text corresponding to the mining geological environment problem instances is segmented, the corresponding expression form keywords and the context containing the expression form keywords are extracted, and the expression form characteristics of each mining geological environment problem are obtained by semantic analysis; the mining geological environment problem instances are clustered according to the expression form characteristics. Since the mining geological environment problem instances contain a large amount of unlabeled data, the mining geological environment problem type label is set for the corresponding clustering cluster according to the type label that accounts for a large proportion in the clustering results, and the causal mechanism of each instance sample is read in each cluster, among which the mining geological environment problem instances are summarized into five categories: land resource occupation and damage, topographic and geomorphic landscape destruction, mining geological disasters, and aquifer destruction; for example, the manifestation of mining geological disasters is the joint action of the air-facing surface and joint fissures formed by open-pit excavation and the original structural surface leading to collapse and sliding disasters; solid waste dumps form landslide disasters under the influence of external factors; landslide disasters caused by ground collapse, etc. The causal mechanisms are integrated, and the correlation between different causal mechanisms and the correlation between the causal mechanisms and the types of mining geological environmental problems are calculated according to the Pearson correlation coefficient. According to two different correlations, indicator factors with high correlation and low redundancy are screened; remote sensing image modal data in mining geological environmental problem instances are extracted, and indicator factors for which remote sensing image modal data cannot be extracted are eliminated. The indicator factors are used to locate in the remote sensing image modal data, and the change representation of the indicator factors of different mining geological environmental problems under the remote sensing image modality is obtained, such as marking the slope cutting and loading range in the development process, the ground subsidence range formed by underground mining, etc.

[0021] Figure 2 A flow chart of extracting mining geological environment elements in a target mining area is shown in an embodiment.

[0022] According to an embodiment of the present invention, the mining geological environment element extraction unit performs mining geological environment element extraction, specifically: S202, the mining geological environment element extraction unit includes a mining geological environment element extraction model, and the mining geological environment element extraction model is constructed based on an improved U-Net, using Res2Net50 to replace the original encoder module; S204, taking the survey area change sequence of the target mining area as the model input, using the hierarchical residual connection in the encoder module Res2Net50 to obtain the feature maps of different scales of the change spots in the survey area change sequence, further downsampling and propagating the encoded feature maps of different scales, and using the residual jump connection as part of the input of the decoder module; S206, in the encoder module, the input feature map is restored by upsampling operation, and the restored feature map of each layer is used as local information to generate prediction branches, and each prediction branch is fused to achieve the integration of global information; S208, the integrated global information is processed by a parallel channel attention module and a spatial attention module to obtain different channel descriptions, the two channel descriptions are added together to achieve fusion, and the fused feature map is identified and extracted from the mining geological environment elements through a Sigmoid layer.

[0023] It should be noted that in the mining geological environment element extraction model, the original encoder module is replaced by Res2Net50. The Res2Net50 network implements hierarchical residual connections in a separate residual block, and uses smaller convolution kernels to more finely represent multi-scale features, enhancing the receptive field range of each network layer. The hierarchical residual connection and multi-scale feature acquisition capabilities are used to comprehensively improve the extraction efficiency and accuracy of mining geological environment elements in the survey area change sequence of the target mining area.

[0024] The feature maps of different scales after encoding are further downsampled and propagated, and residual jump connections are used as part of the input of the decoder module. The decoding blocks of different scales contain both shallow and deep features of the remote sensing image. Each decoding block is restored to the size of the remote sensing image through bilinear upsampling, and used as a prediction branch of the output to achieve feature map integration. The parallel spatial attention module plus the channel attention module explores the dependency between image features. In the channel attention module, the input features are reconstructed to obtain the channel attention map, which is multiplied with the reconstructed features. The result is reconstructed again and added pixel by pixel with the input features to generate the output of the channel attention module, thereby improving the distinguishability of the feature map. The spatial attention module has the same structure as the channel attention module, and the spatial attention module is used to obtain the spatial dependency between any two feature maps at different positions. The decoded output is processed using the parallel channel attention module and the spatial attention module to capture contextual information while giving more weight to the geological environment elements of the mine.

[0025] JaccardLoss is selected as the loss function in the mining geological environment element extraction model. JaccardLoss can focus on pixel-level classification accuracy and spatial overlap at the same time, obtain the geometric and spatial properties of samples, and effectively detect small-area mining geological environment elements.

[0026] It should be noted that the mine geological environment elements in the survey area change sequence are marked, the marked mine geological environment elements are divided into grids, the divided grids are assigned values ​​through the survey area change sequence, and the time series correlation of the grid values ​​is obtained using a time convolution network; the time edges between the grids corresponding to the mine geological environment elements are constructed according to the time series correlation, and the space edges are constructed according to the distance between the grids, and the time and space edges are used to construct a time-space graph for the grids of the mine geological environment elements; the scope of the mine geological environment elements is abstracted into a time-space graph, an undirected graph is used to reduce the amount of calculation, and the maximum influence distance between the grid nodes is preset. If the Euclidean distance between the two grids is less than the maximum influence distance, there is a space edge structure. If it is greater than the maximum influence distance, but can be mutually influenced by another grid node at time t, it is also considered that there is a space edge structure. And according to the time series correlation, it is determined whether there is a time edge, and the grid connected to itself forms a unit matrix, and the matrix directly connected to the spatial edge structure and the matrix directly connected to the time edge structure construct an adjacency matrix. The spatiotemporal graph is represented in the form of an adjacency matrix, the adjacency matrix is ​​learned through a graph neural network, and the features are updated using information transmission and neighbor aggregation mechanisms to obtain the spatiotemporal change characteristics of the mining geological environment elements in the survey area change sequence.

[0027] It should be noted that, in the mining geological environment problem identification unit, the dynamic time warping algorithm is used to calculate the spatiotemporal change characteristics of each mining geological environment element and the change representation of different indicator factors in the remote sensing image mode, and the DTW distance between the feature change sequences is generated; the indicator factor remote sensing image change representation with the minimum DTW distance of each mining geological environment element is compared and obtained, the selected indicator factor remote sensing image change representation is aggregated, and the mining geological environment problems that meet the preset similarity standards are obtained by similarity calculation.

[0028] Figure 3 A flow chart of obtaining the results of a mine geological environment assessment according to an embodiment is shown.

[0029] According to an embodiment of the present invention, the mine geological environment evaluation unit implements the mine geological environment evaluation of the target mine area, specifically: S302, extracting indicator parameters from the mine geological environment problem instance according to the screened indicator factors, quantifying the indicator parameters, obtaining weight information of different indicator factors through the hierarchical analysis method, and performing hierarchical weighted calculation of the weight information indicator factors to obtain the evaluation result of the mine geological environment problem; S304, constructing an evaluation model according to an extreme learning machine, using the mining geological environment problems and their evaluation results as data sources for training and testing the evaluation model, determining the number of neurons in the hidden layer of the model operation, obtaining the accuracy data of the evaluation model, and outputting the evaluation model when the accuracy meets the preset accuracy requirements; S306, extracting the index factors corresponding to the mine geological environment problems of the target mine area, reading the index parameters of the target mine area according to the survey area change sequence of the target mine area based on the index factors, importing them into the evaluation model as model input, and outputting the evaluation prediction results of the target mine area.

[0030] It should be noted that the indicator parameters are quantified. The quantification methods include objective quantification methods such as the range standardization method and the efficacy coefficient method, and subjective quantification methods such as the expert grading method. The weight information of different indicator factors is obtained through the hierarchical analysis method. The weight information indicator factors are graded and weighted to divide the mine geological environment evaluation prediction indicators and comprehensive indexes into different grade intervals, and the corresponding assignment intervals are preset to obtain the evaluation results of mine geological environment problems.

[0031] Another embodiment of the present invention provides a method for evaluating a mine geological environment based on multi-scale features, comprising the following steps: S402, acquiring a remote sensing image of a target mining area as a data source, preprocessing the remote sensing image, and generating a measurement area change sequence of the target mining area; S404, retrieving and acquiring mining geological environmental problems, extracting indicator factors according to the cause mechanism of the mining geological environmental problem examples, and acquiring the change representation of the indicator factors of different mining geological environmental problems under the remote sensing image mode; S406, extracting multi-scale features from the measurement area change sequence for semantic segmentation, obtaining mining geological environment elements, and analyzing the spatiotemporal change characteristics of the mining geological environment elements; S408, performing similarity comparison calculation on the temporal and spatial variation characteristics of the mine geological environment elements and the variation representations of different mine geological environment problems, and identifying the geological environment problems in the target mine area; S410, training an evaluation model based on an extreme learning machine using indicator factors of different mining geological environmental problems, reading indicator parameters of geological environmental problems in a target mining area as model input, and obtaining evaluation and prediction results of the geological environment of the target mining area.

[0032] According to an embodiment of the present invention, a geological environment database of a target mining area is constructed, and evaluation and prediction results of the geological environment of the target mining area are stored in the geological environment database. Warning thresholds for different mining geological environment problems are preset in the geological environment database. If the evaluation result of the geological environment of the target mining area is greater than the corresponding warning threshold, warning information is generated, and environmental restoration data with similarity that meets preset requirements is selected based on historical environmental restoration instances, and the corresponding environmental restoration strategy in the environmental restoration data is extracted as the environmental restoration strategy benchmark for the target mining area. A strategy feasibility analysis is performed based on the geographical characteristics and environmental characteristics of the target mining area, and the environmental restoration strategy for the target mining area is adaptively modified based on the feasibility analysis.

[0033] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as: multiple units or components can be combined, or can be integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the components shown or discussed can be through some interfaces, and the indirect coupling or communication connection of the devices or units can be electrical, mechanical or other forms.

[0034] The units described above as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units; they may be located in one place or distributed on multiple network units; some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0035] In addition, all functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may be separately used as a unit, or two or more units may be integrated into one unit; the above-mentioned integrated units may be implemented in the form of hardware or in the form of hardware plus software functional units.

[0036] A person of ordinary skill in the art can understand that: all or part of the steps of implementing the above method embodiment can be completed by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps of the above method embodiment; and the aforementioned storage medium includes: a mobile storage device, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and other media that can store program codes.

[0037] Alternatively, if the above-mentioned integrated unit of the present invention is implemented in the form of a software function module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiment of the present invention can be essentially or partly reflected in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as mobile storage devices, ROM, RAM, magnetic disks or optical disks.

[0038] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A mine geological environment evaluation system based on multi-scale features, characterized in that: include: Remote sensing image acquisition unit, mining geological environment problem analysis unit, mining geological environment element extraction unit, mining geological environment problem identification unit and mining geological environment evaluation unit; The remote sensing image acquisition unit is responsible for collecting remote sensing images of the target mining area, preprocessing the remote sensing images, and generating a measurement area change sequence of the target mining area; The mining geological environment problem analysis unit is responsible for retrieving mining geological environment problem examples, extracting index factors according to their cause mechanisms, and obtaining the change representation of index factors of different mining geological environment problems under remote sensing image modality; The mining geological environment element extraction unit is responsible for extracting multi-scale features in the measurement area change sequence for semantic segmentation, obtaining mining geological environment elements, and analyzing the temporal and spatial change characteristics of the mining geological environment elements; The mine geological environment problem identification unit is responsible for performing similarity comparison calculation between the temporal and spatial variation characteristics of mine geological environment elements and the variation representations of different mine geological environment problems, and identifying the geological environment problems in the target mine area; The mining geological environment evaluation unit is responsible for training the evaluation model based on the extreme learning machine, reading the index parameters of the geological environment problems in the target mining area as model input, and obtaining the evaluation prediction results of the geological environment of the target mining area.

2. A mine geological environment evaluation system based on multi-scale features according to claim 1, characterized in that: The remote sensing image acquisition unit is specifically: The remote sensing image acquisition unit includes an acquisition unit and a processing unit. The acquisition unit includes a remote sensing imaging device arranged in a satellite. The remote sensing imaging device is used to collect remote sensing image data of a target mining area within a preset time step from a preset remote sensing data source satellite. The processing unit performs image selection, orthophoto processing and image fusion in the remote sensing image data, obtains a high-resolution color image, obtains an aerial photograph of the target mining area as a reference image, crops the high-resolution color image based on the reference image, obtains a satellite image that meets the scope of the mission area, and after completing the remote sensing image preprocessing, generates a remote sensing image sequence according to the acquisition timestamp; In the remote sensing image sequence, the changes of remote sensing images in different periods are analyzed, the change area is marked, the change spots are extracted in the change area, the change spots are interpreted, label information is generated, and a change sequence of spots with label information is obtained; The patch change sequence is aggregated to generate a measurement area change sequence of the target mining area.

3. A mine geological environment evaluation system based on multi-scale features according to claim 1, characterized in that: The mining geological environment problem analysis unit is specifically: Obtaining examples of mining geological environment problems through big data retrieval, segmenting the examples of mining geological environment problems, extracting corresponding expression form keywords and contexts containing expression form keywords, and using semantic analysis to obtain the expression form characteristics of each mining geological environment problem; Clustering the mining geological environment problem instances according to the manifestation characteristics, setting mining geological environment problem type labels for the clustering clusters corresponding to the clustering results, and reading the causal mechanism of each instance sample in each cluster; Integrate the genetic mechanisms, calculate the correlation between different genetic mechanisms and the correlation between the genetic mechanisms and the types of mining geological environmental problems, and select indicator factors according to two different correlations; The remote sensing image modal data in the example of mining geological environment problems are extracted, and the indicator factors are used to locate them in the remote sensing image modal data to obtain the change representation of the indicator factors of different mining geological environment problems under the remote sensing image modality.

4. A mine geological environment evaluation system based on multi-scale features according to claim 1, characterized in that: The mining geological environment element extraction unit is specifically: The mining geological environment element extraction unit includes a mining geological environment element extraction model, which is constructed based on an improved U-Net and uses Res2Net50 to replace the original encoder module; The survey area change sequence of the target mining area is used as the model input. In the encoder module Res2Net50, hierarchical residual connections are used to obtain feature maps of different scales of the change spots in the survey area change sequence. The encoded feature maps of different scales are further downsampled and propagated, and residual jump connections are used as part of the input of the decoder module. In the encoder module, the input feature map is restored through upsampling operation, and the restored feature map of each layer is used as local information to generate prediction branches. Each prediction branch is fused to achieve the integration of global information. The integrated global information is processed by parallel channel attention modules and spatial attention modules to obtain different channel descriptions. The two channel descriptions are added together to achieve fusion, and the fused feature map is identified and extracted from the mining geological environment elements through the Sigmoid layer.

5. A mine geological environment evaluation system based on multi-scale features according to claim 4, characterized in that: In the mining geological environment element extraction unit, the temporal and spatial variation characteristics of the mining geological environment elements are analyzed, specifically: According to the identification and extraction results of the mining geological environment elements in the measurement area change sequence, the marked mining geological environment elements are divided into grids, the divided grids are assigned values ​​through the measurement area change sequence, and the time series correlation of the grid values ​​is obtained using a time convolutional network; Constructing time edges between grids corresponding to mining geological environment elements according to the temporal correlation, and constructing space edges according to the distances between grids, and constructing a space-time graph for the grids of mining geological environment elements using the time edges and space edges; The spatiotemporal graph is represented in the form of an adjacency matrix, the adjacency matrix is ​​learned through a graph neural network, and the features are updated using information transmission and neighbor aggregation mechanisms to obtain the spatiotemporal change characteristics of the mining geological environment elements in the survey area change sequence.

6. A mine geological environment evaluation system based on multi-scale features according to claim 1, characterized in that: The mining geological environment problem identification unit is specifically: The dynamic time warping algorithm is used to calculate the spatiotemporal variation characteristics of each mining geological environment element and the variation representation of different index factors under remote sensing image modality, and the DTW distance between the feature change sequences is generated. The remote sensing image change representations of the indicator factors with the minimum DTW distance of each mining geological environment element are compared, the selected remote sensing image change representations of the indicator factors are aggregated, and the mining geological environment problems that meet the preset similarity standards are obtained by similarity calculation.

7. A mine geological environment evaluation system based on multi-scale features according to claim 1, characterized in that: The mining geological environment assessment unit is specifically: Extracting index parameters from examples of mining geological environment problems based on the selected index factors, quantifying the index parameters, obtaining weight information of different index factors through the hierarchical analysis method, and performing hierarchical weighted calculations on the index factors using the weight information to obtain evaluation results of mining geological environment problems; An evaluation model is constructed based on the extreme learning machine, and the mining geological environment problems and their evaluation results are used as the data source of the evaluation model for training and testing to obtain the accuracy data of the evaluation model. When the accuracy meets the preset accuracy requirements, the evaluation model is output; The index factors corresponding to the mine geological environment problems in the target mine area are extracted, and the index parameters of the target mine area are read according to the survey area change sequence of the target mine area based on the index factors, and are imported into the evaluation model as model input to output the evaluation prediction results of the target mine area.

8. A mine geological environment evaluation method based on multi-scale features, characterized in that: The method is applied to the mine geological environment evaluation system based on multi-scale features as described in any one of claims 1 to 7, comprising the following steps: Acquire a remote sensing image of a target mining area as a data source, pre-process the remote sensing image, and generate a measurement area change sequence of the target mining area; Retrieve and obtain mining geological environmental problems, extract indicator factors according to the causal mechanism of the mining geological environmental problems, and obtain the change representation of the indicator factors of different mining geological environmental problems under the remote sensing image mode; Extracting multi-scale features from the measurement area change sequence for semantic segmentation, obtaining mining geological environment elements, and analyzing the spatiotemporal change characteristics of the mining geological environment elements; Compare and calculate the similarity between the temporal and spatial variation characteristics of mining geological environmental elements and the variation representations of different mining geological environmental problems, and identify the geological environmental problems in the target mining area; Based on the extreme learning machine, the evaluation model is trained using the indicator factors of different mining geological environmental problems. The indicator parameters of the geological environmental problems in the target mining area are read as model input to obtain the evaluation and prediction results of the geological environment of the target mining area.

Images