Satellite remote sensing monitoring method for mining activity of in-situ leaching sandstone type uranium mine

Abstract

The invention belongs to a mining activity monitoring method, and particularly relates to a satellite remote sensing monitoring method for mining activity of an in-situ leaching sandstone type uranium mine, which comprises the following steps of: 1, acquiring and preprocessing remote sensing data; 2, preliminarily screening mining active areas based on soil spectral characteristics; step 3, well site range fine extraction fused with image texture features; 4, identifying a central control room based on morphological characteristics; 5, comprehensively judging the range of the well site; and 6, mining activity dynamic monitoring. According to the method, the unique remote sensing features of the in-situ leaching sandstone type uranium mine mining activity are fully excavated, accurate recognition and dynamic change monitoring of the mine mining area are achieved, and efficient and reliable technical support is provided for monitoring of the in-situ leaching sandstone type uranium mine mining activity.

Description

Technical Field

[0001] This invention pertains to methods for monitoring mining activities, specifically a satellite remote sensing method for monitoring mining activities in in-situ leaching sandstone-type uranium mines. Background Technology

[0002] Remote sensing technology, with its advantages of high resolution, strong timeliness and wide coverage, can play an important role in the monitoring of mining activities. Its current main applications include: (1) using multi-temporal high-resolution optical images to monitor land use changes in mining areas and analyze mining activities; (2) using radar images to extract deformation data of mining areas and monitor changes in mining subsidence areas; (3) using a landscape ecological assessment model based on the fusion of remote sensing and GIS to assess the ecological quality and changes of mining areas; and (4) using thermal infrared remote sensing monitoring technology to monitor thermal activity information in ore piles and tailings ponds.

[0003] The above-mentioned remote sensing technologies are mainly used for monitoring activities in heap leaching mines. However, in-situ leaching sandstone uranium deposits are a relatively special type of mine. The mining method involves injecting leaching solution into the ore layer through injection holes at the surface, leaching in situ underground, and then extracting the ore-bearing solution through extraction holes for refining. No ore extraction is required. Therefore, there are fewer traces of surface human activity in in-situ leaching sandstone uranium mines, which limits the effectiveness of remote sensing technology in monitoring their mining activities. Currently, remote sensing technology is mainly used for studying the geological background and metallogenic theories of in-situ leaching sandstone uranium deposits, with limited research on monitoring their mining activities.

[0004] Based on the problems encountered by remote sensing technology in monitoring mining activities in in-situ leaching sandstone uranium mines, it is necessary to establish a method to more effectively extract mining activity information, thereby improving the effectiveness of remote sensing monitoring. Summary of the Invention

[0005] The purpose of this invention is to provide a satellite remote sensing monitoring method for in-situ leaching sandstone uranium mining activities. This method fully utilizes the unique remote sensing characteristics of in-situ leaching sandstone uranium mining activities to achieve accurate identification and dynamic change monitoring of mining areas, providing efficient and reliable technical support for monitoring in-situ leaching sandstone uranium mining activities.

[0006] Technical solution to achieve the purpose of this invention:

[0007] A satellite remote sensing monitoring method for in-situ leaching sandstone-type uranium mine operations includes:

[0008] Step 1: Remote sensing data acquisition and preprocessing;

[0009] Step 2: Preliminary screening of mining activity areas based on soil spectral characteristics;

[0010] Step 3: Fine extraction of the well site range by fusing image texture features;

[0011] Step 4: Control room identification based on morphological features;

[0012] Step 5: Comprehensive determination of the well site area;

[0013] Step Six: Dynamic Monitoring of Mining Activities.

[0014] Furthermore, step one includes: acquiring multi-temporal satellite remote sensing image data of the monitoring area, with the image band range covering the visible light to shortwave infrared range, the number of bands being at least 8, and the spatial resolution being better than 0.5 meters; preprocessing the acquired remote sensing image data, the preprocessing operations including radiometric correction, geometric correction, atmospheric correction, image mosaicking, and image cropping, to obtain a standardized remote sensing image dataset.

[0015] Further, step two includes: based on images of known mining activity areas, uniformly selecting soil endmembers from the mining activity areas and the surrounding native areas on the images to construct a spectral library of soil endmembers from the mining activity areas and the surrounding native areas; based on the constructed endmember spectral library, classifying soil types using a spectral angle algorithm with a spectral angle threshold of 0.15, filtering mining activity areas on the standardized remote sensing image dataset obtained in step one, and obtaining the soil extent of the mining activity areas and the surrounding native areas in the entire image to obtain remote sensing images of the mining activity areas.

[0016] Furthermore, step three includes:

[0017] Step 3.1: Based on the remote sensing images of the mining area extracted in Step 2, extract the texture features of the images using the gray-level co-occurrence matrix algorithm, and combine the texture features with the image's own bands to construct the "spectral-texture" fusion feature data of the target area.

[0018] Step 3.2: Based on known well site images, pixel information of the three target well site ranges (soil, pumping holes, and injection holes) within the mining area is used as positive samples, and pixel information of other non-well site ranges within the mining area is used as negative samples. Texture features are extracted from the known well site images and combined with their own bands to generate "spectral-texture" fusion feature data of the sample area. The pixel coordinates of the positive and negative samples are matched with their "spectral-texture" fusion feature data to form a paired standard training dataset. The paired standard training dataset is input into a random forest classifier for training and classification to obtain a trained well site recognition model.

[0019] Step 3.3: Input the constructed "spectral-texture" fusion feature data of the target area into the trained well site recognition model, and output raster data containing both well site and non-well site targets.

[0020] Furthermore, the texture features in step three include contrast, correlation, energy, and entropy.

[0021] Furthermore, step four includes: based on the remote sensing image of the mining area obtained in step two, using an edge detection algorithm to extract building edge information, combining the preset control room outline regularity and actual length and width features, constructing a control room identification model and screening control rooms, and outputting the extracted control room raster data.

[0022] Furthermore, the preset outline regularity and actual length and width characteristics of the central control room in step four are: outline regularity ≥ 0.8, actual length 5-20m, and width 3-10m.

[0023] Furthermore, step five includes:

[0024] Step 5.1, Grid Data Assignment: For the grid data output in Step 3, which includes both well site and non-well site targets, assign a value of 1 to the well site grid data and a value of 0 to the non-well site grid data. This data layer is named the Well Site Basic Layer. For the control room grid data output in Step 4, assign a value of 1 to the control room grid data. Then, perform interpolation and normalization on the entire data. Finally, the values ​​of all control room grids are within the range of 0 to 1. This data layer is named the Control Room Association Layer.

[0025] Step 5.2, Raster Data Fusion Calculation: The two layers of raster data after the above assignment are superimposed using a 1:1 equal weight. The score is calculated according to the comprehensive score calculation formula: Comprehensive Score = (Well Site Basic Layer Value + Control Room Related Layer Value) / 2. Based on the score, well site candidate areas are divided, and small isolated patches in the well site candidate areas are removed. The raster data of the remaining connected areas are converted into vector area data, the boundary smoothness is optimized, and an accurate well site range vector file is output.

[0026] Furthermore, in step 5.2, grids with a score ≥ 0.55 are divided into well site candidate regions, and areas ≤ 9m² are removed. 2 Isolated small patches.

[0027] Furthermore, step six includes: acquiring multi-temporal remote sensing data of the monitored mine according to the time period, repeating steps one to five for each temporal data to obtain vector data of the well site range for each temporal phase; using the initial data as a benchmark, calculating the change rate of well site area and spatial transfer in each phase through overlay analysis, summarizing the change pattern of mining activities, and realizing dynamic tracking.

[0028] The beneficial technical effects of this invention are as follows:

[0029] 1. The present invention provides a satellite remote sensing monitoring method for mining activities in in-situ leaching sandstone uranium mines, which can effectively improve monitoring efficiency: applying satellite remote sensing technology to the monitoring of mining activities in in-situ leaching sandstone uranium mines can make full use of the advantages of remote sensing technology, such as wide coverage, high resolution, and strong timeliness, to achieve rapid monitoring of large areas and improve monitoring efficiency.

[0030] 2. The present invention provides a satellite remote sensing monitoring method for in-situ leaching sandstone type uranium mine mining activities, which can effectively improve monitoring accuracy: targeting the unique engineering characteristics of in-situ leaching sandstone type uranium mine mining, key information is extracted from three dimensions: image texture features (injection holes / extraction holes), spectral features (soil in the mining area), and morphological features (central control room). Through multi-dimensional feature comprehensive analysis, the mining area range can be accurately determined, effectively solving the problem of insufficient accuracy of single feature recognition. Attached Figure Description

[0031] Figure 1 The flowchart illustrates a satellite remote sensing monitoring method for in-situ leaching sandstone-type uranium mine operations provided by this invention. Detailed Implementation

[0032] The present invention will now be described in further detail with reference to the accompanying drawings and embodiments.

[0033] like Figure 1 As shown, this invention provides a satellite remote sensing monitoring method for in-situ leaching sandstone-type uranium mining activities, specifically including the following steps:

[0034] Step 1: Remote Sensing Data Acquisition and Preprocessing

[0035] Acquire multi-temporal satellite remote sensing image data of the monitoring area, covering the visible to shortwave infrared band range, with at least 8 bands and a spatial resolution better than 0.5 meters. Preprocess the acquired remote sensing image data, including radiometric correction, geometric correction, atmospheric correction, image mosaicking, and image cropping, to obtain a standardized remote sensing image dataset.

[0036] Step 2: Preliminary screening of mining activity areas based on soil spectral characteristics

[0037] Mining activities alter the shape, color, and other characteristics of the soil. As a result, there are differences in the spectral characteristics of the soil in the mining area and the surrounding original area. Based on these differences, the extent of mining activities in the mining area can be preliminarily determined.

[0038] First, based on images of known mining areas, soil endmembers from both the mining areas and the surrounding native areas are uniformly selected from the images to construct spectral libraries for these endmembers. Then, based on these constructed endmember spectral libraries, soil type classification is performed using a spectral angle algorithm with a spectral angle threshold of 0.15. The standardized remote sensing image dataset obtained in step one is then used to filter for mining areas, obtaining the soil extent of both the mining areas and the surrounding native areas within the entire image, thus generating remote sensing images of the mining areas.

[0039] Step 3: Fine extraction of well site range by fusing image texture features

[0040] In in-situ leaching sandstone-type uranium mines, the well site (the distribution area of ​​injection and extraction wells) is the actual mining area. Besides the well site, there are also artificial facilities such as office areas and processing workshops, as well as areas for human activities such as cargo transportation. Therefore, the mining activity area extracted in step two includes the well site area and other areas of human activity. The core of using remote sensing technology to monitor mining activities in in-situ leaching sandstone-type uranium mines is monitoring changes in the well site's extent; therefore, it is necessary to further extract the well site extent based on step two. Injection and extraction wells are mostly circular in shape, with diameters typically ranging from 50 cm to 1 meter. While it is difficult to directly extract injection and extraction wells from satellite remote sensing images, they significantly affect the image's texture features. Therefore, the well site extent is indirectly extracted based on texture features.

[0041] Step 3.1: Based on the remote sensing images of the mining area extracted in Step 2, use the gray-level co-occurrence matrix algorithm to extract four types of texture features of the image: contrast, correlation, energy, and entropy. Combine the four types of texture features with the image's own bands to construct the "spectral-texture" fusion feature data of the target area.

[0042] Step 3.2: Based on known well site images, pixel information of the three target well site ranges (soil, pumping holes, and injection holes) within the mining area is used as positive samples, and pixel information of other non-well site ranges within the mining area is used as negative samples. Texture features are extracted from the known well site images and combined with their own bands to generate "spectral-texture" fusion feature data of the sample area. The pixel coordinates of the positive and negative samples are matched with their "spectral-texture" fusion feature data to form a paired standard training dataset. The paired standard training dataset is input into a random forest classifier for training and classification to obtain a trained well site recognition model.

[0043] Step 3.3: Input the "spectral-texture" fusion feature data of the target area constructed in Step 3.1 into the well site recognition model trained in Step 3.2, and output raster data containing both well site and non-well site targets.

[0044] Step 4: Control Room Identification Based on Morphological Features

[0045] The control room is used to control the mining activities of injection and extraction holes within a certain space. There are several control rooms within the well site.

[0046] Based on the remote sensing images of the mining area obtained in step two, the Canny edge detection algorithm is used to extract building edge information. Combined with the preset characteristics of the central control room, namely "outline regularity ≥ 0.8, actual length 5-20m, width 3-10m", a central control room identification model is constructed and central control rooms are screened, and the extracted central control room raster data is output.

[0047] Step 5: Comprehensive Determination of Well Site Scope

[0048] To more accurately delineate the well site area, it is necessary to conduct a comprehensive analysis using the well site grid data extracted in step three and the central control room grid data extracted in step four.

[0049] Step 5.1: Assigning raster data

[0050] For the grid data output in step three, which includes both well site and non-well site targets, the well site grid data is assigned a value of 1, and the non-well site grid data is assigned a value of 0. This data layer is named the well site basic layer. For the control room grid data output in step four, the control room grid data is assigned a value of 1. Then, the entire data is interpolated and normalized. Finally, the values ​​of all control room grids are in the range of 0 to 1. The larger the value, the closer the spatial distance to the control room. This data layer is named the control room association layer.

[0051] Step 5.2: Raster data fusion calculation

[0052] The two layers of raster data, after being assigned values, were superimposed using a 1:1 equal weighting. The score was calculated according to the comprehensive score formula: Comprehensive Score = (Well Site Base Layer Value + Control Room Related Layer Value) / 2. Grids with a score ≥ 0.55 were designated as well site candidate areas, and areas ≤ 9m² were removed. 2 The isolated small patches (excluding data noise interference) are then processed, and the remaining connected region's raster data is converted into vector area data. The boundary smoothness is optimized, and an accurate well site range vector file is output.

[0053] Step Six: Dynamic Monitoring of Mining Activities

[0054] Acquire multi-temporal remote sensing data of the monitored mine at a certain time period. Repeat steps one to five for each temporal data to obtain vector data of the well site range for each temporal phase. Using the initial data as a benchmark, calculate the change rate of well site area and spatial transfer in each phase through overlay analysis, summarize the change pattern of mining activities, and achieve dynamic tracking.

[0055] Example 1

[0056] A typical in-situ leached sandstone uranium mine was selected as the monitoring target area, and satellite remote sensing monitoring of mining activities in the in-situ leached sandstone uranium mine was carried out. The specific steps included:

[0057] Step 1: Remote Sensing Data Acquisition and Preprocessing

[0058] High-resolution optical remote sensing images of the monitored area were acquired. The image type was WorldView-3 satellite data, with a spatial resolution of 0.3m, a total of 16 bands, and 3 scenes. The acquisition times were January 2020, June 2022, and August 2024, respectively.

[0059] The remote sensing images were preprocessed as follows: radiometric correction was performed using ENVI software to eliminate sensor errors; geometric correction was performed using ground control points to ensure that the spatial position accuracy error of the image was less than 1 pixel; atmospheric effects were removed using the FLAASH atmospheric correction model; and the images were cropped according to the boundary of the monitoring target area to obtain a standardized remote sensing image dataset.

[0060] Step 2: Preliminary screening of mining activity areas based on soil spectral characteristics

[0061] First, based on images of known mining areas, soil endmembers from both the mining area and the surrounding native area are uniformly selected from the images. Soil samples (30 from the mining area and 30 from the surrounding native area) are collected separately to construct endmember spectral libraries for both areas. Then, based on these endmember spectral libraries, soil type classification is performed using a spectral angle algorithm with a threshold of 0.15. The standardized remote sensing image dataset obtained in step one is then used to filter for mining areas, obtaining the soil extent of both the mining area and the surrounding native area within the entire image, thus generating remote sensing images of the mining areas.

[0062] Step 3: Fine extraction of well site range by fusing image texture features

[0063] Using the remote sensing images of the mining area extracted in step two, the gray-level co-occurrence matrix algorithm was used to extract four types of texture features: contrast, correlation, energy, and entropy. These four texture features were then combined with the image's own 16 bands to construct "spectral-texture" fused feature data, totaling 20 bands. Based on images of known well sites, pixel information of three types of targets within the mining area—soil, pumping holes, and injection holes—was used as positive samples (60 pixels), while pixel information of other non-well site areas within the mining area was used as negative samples (50 pixels). These samples were then input into a random forest classification model for training and classification. The model had 100 decision trees and output raster data containing both well site and non-well site targets.

[0064] Step 4: Control Room Identification Based on Morphological Features

[0065] Based on the remote sensing images of the mining area obtained in step two, after denoising with 3×3 Gaussian filtering and linear stretching enhancement, the Canny edge detection algorithm (high threshold 0.2, low threshold 0.05) is used to extract building edge information. Combined with the characteristics of the central control room, namely "outline regularity ≥ 0.8, actual length 5-20m, width 3-10m", a central control room identification model is constructed and central control rooms are screened, and the extracted central control room raster data is output.

[0066] Step 5: Comprehensive Determination of Well Site Scope

[0067] First, grid values ​​are assigned. For the output data from step three, well site grid data is assigned a value of 1, and non-well site grid data is assigned a value of 0. This data layer is named the well site basic layer. For the control room grid data output from step four, the control room grid data is assigned a value of 1. Then, interpolation and normalization are performed on the entire data. Finally, the values ​​of all control room grids are in the range of 0 to 1. The larger the value, the closer the spatial distance to the control room. This data layer is named the control room association layer.

[0068] Then, a fusion calculation is performed. The two layers of raster data, after being assigned values, are superimposed using a 1:1 equal weighting. The comprehensive score is calculated as follows: Comprehensive Score = (Well Site Base Layer Value + Control Room Related Layer Value) / 2. Grids with a score ≥ 0.55 are designated as well site candidate areas, and areas ≤ 9m² are removed. 2 The isolated small patches (excluding data noise interference) are then processed, and the remaining connected region's raster data is converted into vector area data. The boundary smoothness is optimized, and an accurate well site range vector file is output.

[0069] Step Six: Dynamic Monitoring of Mining Activities

[0070] Steps one through five were repeated for three remote sensing data points from different time periods to obtain vector data of the well site area for each time period. The well site area in January 2020 was 16.32 km². 2 The well site area in June 2022 was 18.68 km². 2 The well site area in August 2024 was 23.18 km². 2 Based on the initial data, the change rate of well site area and spatial transfer in each period were calculated through overlay analysis. The analysis showed that the well site area has been expanding eastward, and the change rate of well site area is consistent with the mine's production capacity.

[0071] The present invention has been described in detail above with reference to the accompanying drawings and embodiments. However, the present invention is not limited to the above embodiments, and various changes can be made within the scope of knowledge possessed by those skilled in the art without departing from the spirit of the present invention. All contents not described in detail in the present invention can be derived from existing technologies.

Claims

1. A satellite remote sensing monitoring method for mining activities in in-situ leaching sandstone type uranium mines, characterized in that, include: Step 1: Remote sensing data acquisition and preprocessing; Step 2: Preliminary screening of mining activity areas based on soil spectral characteristics; Step 3: Fine extraction of the well site range by fusing image texture features; Step 4: Control room identification based on morphological features; Step 5: Comprehensive determination of the well site area; Step Six: Dynamic Monitoring of Mining Activities.

2. The satellite remote sensing monitoring method for in-situ leaching sandstone-type uranium mine mining activities according to claim 1, characterized in that, Step one includes: acquiring multi-temporal satellite remote sensing image data of the monitoring area, with the image band range covering the visible light to shortwave infrared range, the number of bands being at least 8, and the spatial resolution being better than 0.5 meters; preprocessing the acquired remote sensing image data, the preprocessing operations including radiometric correction, geometric correction, atmospheric correction, image mosaicking, and image cropping, to obtain a standardized remote sensing image dataset.

3. The satellite remote sensing monitoring method for in-situ leaching sandstone-type uranium mine mining activities according to claim 2, characterized in that, Step two includes: based on images of known mining areas, uniformly selecting soil endmembers from the mining areas and surrounding native areas on the images to construct a spectral library of soil endmembers from the mining areas and surrounding native areas; based on the constructed endmember spectral library, classifying soil types using a spectral angle algorithm with a spectral angle threshold of 0.15, filtering mining areas on the standardized remote sensing image dataset obtained in step one, obtaining the soil extent of the mining areas and surrounding native areas in the entire image, and obtaining remote sensing images of the mining areas.

4. The satellite remote sensing monitoring method for in-situ leaching sandstone-type uranium mine mining activities according to claim 3, characterized in that, Step three includes: Step 3.1: Based on the remote sensing images of the mining area extracted in Step 2, extract the texture features of the images using the gray-level co-occurrence matrix algorithm, and combine the texture features with the image's own bands to construct the "spectral-texture" fusion feature data of the target area; Step 3.2: Based on known well site images, pixel information of the three target well site ranges (soil, pumping holes, and injection holes) within the mining area is used as positive samples, and pixel information of other non-well site ranges within the mining area is used as negative samples. Texture features are extracted from the known well site images and combined with their own bands to generate "spectral-texture" fusion feature data of the sample area. The pixel coordinates of the positive and negative samples are matched with their "spectral-texture" fusion feature data to form a paired standard training dataset. The paired standard training dataset is input into a random forest classifier for training and classification to obtain a trained well site recognition model. Step 3.3: Input the constructed "spectral-texture" fusion feature data of the target area into the trained well site recognition model, and output raster data containing both well site and non-well site targets.

5. A satellite remote sensing monitoring method for in-situ leaching sandstone-type uranium mine mining activities according to claim 4, characterized in that, The texture features in step three include contrast, correlation, energy, and entropy.

6. A satellite remote sensing monitoring method for in-situ leaching sandstone-type uranium mine mining activities according to claim 4, characterized in that, Step four includes: based on the remote sensing image of the mining area obtained in step two, using an edge detection algorithm to extract building edge information, combining the preset control room outline regularity and actual length and width features, constructing a control room identification model and screening control rooms, and outputting the extracted control room raster data.

7. A satellite remote sensing monitoring method for in-situ leaching sandstone-type uranium mine mining activities according to claim 6, characterized in that, The preset outline regularity and actual length and width characteristics of the central control room in step four are: outline regularity ≥ 0.8, actual length 5-20m, and width 3-10m.

8. A satellite remote sensing monitoring method for in-situ leaching sandstone-type uranium mine mining activities according to claim 6, characterized in that, Step five includes: Step 5.1, Grid Data Assignment: For the grid data output in Step 3, which includes both well site and non-well site targets, assign a value of 1 to the well site grid data and a value of 0 to the non-well site grid data. This data layer is named the Well Site Basic Layer. For the control room grid data output in Step 4, assign a value of 1 to the control room grid data. Then, perform interpolation and normalization on the entire data. Finally, the values ​​of all control room grids are within the range of 0 to 1. This data layer is named the Control Room Association Layer. Step 5.2, Raster Data Fusion Calculation: The two layers of raster data after the above assignment are superimposed using a 1:1 equal weight. The score is calculated according to the comprehensive score calculation formula: Comprehensive Score = (Well Site Basic Layer Value + Control Room Related Layer Value) / 2. Based on the score, well site candidate areas are divided, and small isolated patches in the well site candidate areas are removed. The raster data of the remaining connected areas are converted into vector area data, the boundary smoothness is optimized, and an accurate well site range vector file is output.

9. A satellite remote sensing monitoring method for in-situ leaching sandstone-type uranium mine mining activities according to claim 8, characterized in that, In step 5.2, grid cells with a score ≥ 0.55 are classified as well site candidate regions, and areas ≤ 9m² are removed. 2 Isolated small patches.

10. A satellite remote sensing monitoring method for in-situ leaching sandstone-type uranium mine mining activities according to claim 8, characterized in that, Step six includes: acquiring multi-temporal remote sensing data of the monitored mine according to time periods; repeating steps one to five for each temporal data to obtain vector data of the well site range for each temporal phase; using the initial data as a benchmark, calculating the change rate of well site area and spatial transfer in each phase through overlay analysis, summarizing the changing patterns of mining activities, and achieving dynamic tracking.

Images