Mine ecological environment monitoring method and system fusing soil heavy metal ecological indexes

Abstract

The invention discloses a mine ecological environment monitoring method and system fusing soil heavy metal ecological indexes, and the method comprises the steps: calculating a corresponding potential ecological risk index RI based on the soil heavy metal content of a sample point of a mine sample region, and carrying out the fitting of remote sensing data of the sample point and the RI, and obtaining an RI inversion model; performing wave band operation on the multispectral remote sensing data of the mine sample area, and screening and constructing an optimal wave band combination; using the optimal wave band of the sample region and RI-ML regression to obtain an RRI inversion model; using the RRI inversion model and according to the optimal wave band of the mine monitoring area, performing inversion on the mine monitoring area to obtain a remote sensing ecological risk index RRI; the method comprises the following steps: calculating NDVI, WET and NDBSI by using multispectral remote sensing data of a mine monitoring area, then calculating a first principal component as a pollution enhanced remote sensing ecological index PE-RSEI, and carrying out space-time comparison on the ecological environment of the mine monitoring area; according to the method, simple and feasible soil heavy metal pollution fusion indexes are constructed and fused into a mining area ecological monitoring system, the current situation of mining area ecological environment quality can be visually reflected, and reliable reference is provided for mining area ecological monitoring and protection work.

Description

Technical Field

[0001] This invention belongs to the field of ecological monitoring and remote sensing technology, specifically relating to a method and system for monitoring the ecological environment of mines that integrates soil heavy metal ecological indicators. Background Technology

[0002] Traditional integrated ecological and environmental monitoring methods primarily rely on ecological and environmental monitoring stations and field sampling. These stations utilize physical observations or environmental geochemical methods to acquire discrete data on ecological and environmental quality, which is then used to obtain isometric ecological and environmental quality data through spatial interpolation. The advantages of this method are high accuracy and a strong correlation between monitoring station results and local conditions. However, its disadvantages include the significant resource consumption required for station deployment, difficulty in retrospective monitoring, and challenges in long-term monitoring and time-series analysis. Furthermore, spatial interpolation methods suffer from insufficient accuracy and applicability in the inversion of spatial data at medium scales.

[0003] Spaceborne remote sensing, with its large scale, periodicity, and relatively high precision, has gradually become a common means of ecological environment monitoring. Traditional ecological environment monitoring based on spaceborne remote sensing technology can be mainly divided into the following categories: (1) Ecological monitoring based on single bands. Some vegetation and land cover types are highly sensitive to near-infrared and short-infrared bands, and the changes in single bands can reflect the characteristics of changes in the surface ecological environment; (2) Remote sensing index method. The surface features are extracted through multi-band calculations to reflect the quality of the surface ecological environment. Common ones include Normalized Difference Vegetation Index (NDVI), Normalized Difference Water Index (NDWI), Enhanced Vegetation Index (EVI), etc. A single remote sensing index can only characterize the change of a certain element on the surface and does not have the performance of comprehensive ecological environment inversion; (3) Comprehensive model method. A comprehensive model is constructed by integrating multiple remote sensing indices or other natural elements to comprehensively invert changes in the surface ecological environment. Common ones include Remote Sensing Ecological Indicator (RSEI) and Habitat Quality Model (InVest), etc. However, common comprehensive models hardly involve the relevant content of soil heavy metal pollution risk.

[0004] Regarding the ecological environment of mines, there are currently no effective solutions to questions such as how to effectively identify the ecological risks of soil heavy metal pollution, how to achieve historical inversion and time series analysis of the ecological risks of soil heavy metal pollution, and how to integrate the ecological risks of soil heavy metal pollution into the comprehensive model of ecological environment quality assessment. Summary of the Invention

[0005] This invention provides a method and system for monitoring the ecological environment of mines that integrates soil heavy metal ecological indicators, which can improve monitoring accuracy.

[0006] To achieve the above technical objectives, the present invention adopts the following technical solution:

[0007] A method for monitoring the ecological environment of mines that integrates soil heavy metal ecological indicators includes:

[0008] The potential ecological risk index (RI) is calculated based on the soil heavy metal content of sample points in the mining area, and the RI inversion model is obtained by regression fitting using multispectral remote sensing data of the sample points and RI.

[0009] Band calculations were performed on the multispectral remote sensing data of the mining sample area, and the optimal band combination was selected and constructed based on the calculated bands and the RI-ML obtained by the RI inversion model of the original bands.

[0010] The RRI inversion model is obtained by using the optimal band combination data of the mine sample area and RI-ML regression fitting.

[0011] The Remote Sensing Ecological Risk Index (RRI) is obtained by using the RRI inversion model and based on the optimal combination of remote sensing bands in the mine monitoring area.

[0012] The greenness NDVI, humidity WET, and dryness NDBSI of the mining monitoring area were calculated using multispectral remote sensing data. Then, principal component analysis was used to calculate RRI, NDVI, WET, and NDBSI, and the first principal component was used as the pollution enhancement remote sensing ecological index PE-RSEI.

[0013] Spatiotemporal analysis of the ecological environment in the mine monitoring area was conducted based on PE-RSEI.

[0014] Furthermore, the calculation method for the potential ecological risk index RI is as follows:

[0015] ,

[0016] In the formula, It is the first Potential ecological risk index for each sample point It is the first The first sample point Potential ecological risks of heavy metals; It is the first The toxicity response coefficient of a heavy metal reflects its toxicity to the human body and its sensitivity to environmental migration. It is the first At the nth sample point The content of various heavy metals; It is the first At the nth sample point Reference values ​​for the upper limit of pollution of certain heavy metals.

[0017] Furthermore, six original bands of data from multispectral remote sensing data—visible, near-infrared, and short-infrared—were selected as independent variables, and the natural logarithm of the potential ecological risk index (RI) was used as the dependent variable. The random forest method was then used to perform regression fitting on the independent and dependent variables to obtain the RI inversion model.

[0018] Furthermore, performing band calculations, filtering, and constructing optimal band combinations on multispectral remote sensing data includes:

[0019] Six original bands of data from multispectral remote sensing data were selected, including visible light, near infrared, and short infrared bands. One, two, and three bands were randomly selected from these bands to perform one-dimensional, two-dimensional, and three-dimensional operations, resulting in a total of 738 candidate band data.

[0020] Using each candidate band data as the independent variable and the RI-ML obtained from the original band via the RI inversion model as the dependent variable, the correlation between the independent and dependent variables and the multicollinearity between the independent and dependent variables are calculated using a random sampling method. N bands with multicollinearity less than the preset value and the highest correlation with the dependent variable are selected from all candidate bands.

[0021] The selected N bands are arranged and combined according to the pre-selected quantity; then, each combination is used as a set of independent variables, and RI-ML is used as the dependent variable for linear regression. The optimal band combination is selected based on the root mean square error (RSME) of the linear regression model corresponding to each set of independent variables.

[0022] Furthermore, the least partial squares regression method was used, and the band combination data of the sample points and RI were used for regression fitting to obtain the corresponding linear regression model.

[0023] Furthermore, the optimal band combinations include: , , , ;in, The values ​​represent the reflectance of the red band, green band, blue band, near-infrared band, short-infrared 1 band, and short-infrared band 2 bands obtained from the Landsat satellite multispectral sensor. , , , These represent the functions for square root calculation, natural logarithm calculation, predefined scheme calculation, and ratio calculation, respectively, and are expressed as follows:

[0024]

[0025]

[0026]

[0027] .

[0028] Furthermore, before using principal component analysis to calculate RRI, NDVI, WET, and NDBSI, the data of each remote sensing ecological risk index (RRI), greenness (NDVI), humidity (WET), and dryness (NDBSI) are normalized. Then, principal component analysis is used to calculate the principal components of the normalized RRI, NDVI, WET, and NDBSI.

[0029] Furthermore, the spatiotemporal analysis of the ecological environment of the mine monitoring area based on PE-RSEI includes: the larger the PE-RSEI of a certain period, the higher the ecological environment quality of that period; the larger the PE-RSEI of a certain location point, the higher the ecological environment quality of that location.

[0030] A mine ecological environment monitoring system integrating soil heavy metal ecological indicators includes:

[0031] The RI calculation module is used to calculate the corresponding potential ecological risk index (RI) based on the soil heavy metal content of sample points in the mining sample area.

[0032] The RI fitting module is used to: perform regression fitting between multispectral remote sensing data of sample points and RI to obtain an RI inversion model;

[0033] The band calculation and filtering module is used to: perform band calculations on multispectral remote sensing data of the mine sample area, and filter and construct the best band combination based on the calculated bands and the RI-ML obtained by the RI inversion model of the original bands.

[0034] The RRI fitting module is used to obtain an RRI inversion model by performing regression fitting using the best band combination data of the mine sample area and RI-ML.

[0035] The RRI inversion module is used to: use the RRI inversion model and based on the optimal combination of remote sensing bands in the mining monitoring area to invert the remote sensing ecological risk index (RRI) of the mining monitoring area.

[0036] The principal component analysis module is used to: calculate the greenness NDVI, humidity WET, and dryness NDBSI of the mine monitoring area using multispectral remote sensing data, and then use principal component analysis to calculate RRI, NDVI, WET, and NDBSI, and extract the first principal component as the pollution enhancement remote sensing ecological index PE-RSEI.

[0037] The monitoring module is used for: spatiotemporal analysis of the ecological environment of the mine monitoring area based on PE-RSEI.

[0038] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0039] To address the issue that commonly used remote sensing ecological indicators (RSEIs) in current environmental monitoring lack soil pollution components and cannot be applied to mining areas, this invention proposes a method for assessing the ecological environment quality of mining areas based on remote sensing ecological technology. This invention selects surface heavy metal pollution data to construct a remote sensing surface soil heavy metal ecological risk model (RRI), and integrates it into the RSEI system to build a pollution-enhanced remote sensing ecological indicator system, namely the PE-RSEI indicator system. This PE-RSEI indicator system can then be used to simply and intuitively assess the current status of the ecological environment quality of mining areas, facilitating widespread application and reducing the cost of ecological environment monitoring. Furthermore, this method can utilize the PE-RSEI indicator to conduct spatiotemporal monitoring and analysis of mine ecology, completing time-series monitoring of the ecological environment of mining areas. Attached Figure Description

[0040] Figure 1 This is a framework diagram of the method described in the embodiments of this application.

[0041] Figure 2 This application embodiment shows the spatial distribution of the potential ecological risk index RI, single-phase RI, and RRI in the mining monitoring area in 2024. Each sub-figure (a), (b), and (c) corresponds to RI, RI-ML, and RRI, respectively.

[0042] Figure 3 The spatial distribution of PE-RSEI for each year from 1997 to 2024 is shown in the figure. Figure 3 The sub-figures are numbered (1997) to (2024) respectively, corresponding to the years. Detailed Implementation

[0043] The embodiments of the present invention will be described in detail below. These embodiments are based on the technical solutions of the present invention and provide detailed implementation methods and specific operation processes to further explain the technical solutions of the present invention.

[0044] Example 1

[0045] This embodiment provides a method for monitoring the ecological environment of mines by integrating soil heavy metal ecological indicators. The satellite images used are mainly Landsat 5, 7, 8, 9 and Sentinel-2 multispectral satellite data (Table 1). Among them, Landsat data is used to invert the band calculation formula to realize the time series analysis of surface pollution. The main preprocessing process of multispectral satellite images includes: (1) Image screening: All available images are screened according to the location of the study area, cloud cover and satellite product quality. Landsat 7 satellite has stripe failure, so the use of Landsat 7 images is minimized. (2) Image declouding and band selection. (3) Image synthesis: All available images for each year are used to synthesize one image per year using the median synthesis method (Table 1) to form an image set from 1997 to 2024. Imagex represents the median synthesized image of the year in the dataset. Image2024 is used together with the heavy metal content data from field tests to invert RI, so as to obtain a more accurate spatial distribution of RI instead of the traditional spatial interpolation method. Sentinel-2 images have high spatial resolution and are used for visualization and auxiliary calculation. All of the above image collection and preprocessing work was completed in Google Earth Engine (GEE).

[0046]

[0047] Based on the industrial production characteristics of the study area, a field sampling experiment was designed, with the main objective of collecting soil samples to test the content of relevant heavy metals. Sampling took place in October 2024, and the spatial distribution of sampling points is shown below. Figure 1 The soil sampling principle was as follows: sampling points were evenly distributed around typical areas such as mining areas, roads, and catchment areas, taking into account accessibility. At each sampling point, 200g of topsoil (0-20 cm) was collected from each of the four corners and the center of a 20m² quadrat using a 5-point sampling method. The samples were evenly mixed and placed in clean plastic bags. The geographical coordinates of the sampling point center were obtained using a GPS locator, and the sampling time, sampling point number, and soil type were marked. A total of 94 samples were obtained. Figure 1 Each sample was air-dried, ground, and passed through a 100-mesh nylon sieve in the laboratory before being used for the next step of processing.

[0048] Based on the current production status of the study area, all soil samples were used to determine the contents of As, Cd, Cu, Hg, Ni, Zn, and Pb, using methods according to the "Regional Geochemical Sample Analysis Methods" issued by the Ministry of Natural Resources of China. Soil samples were digested with HNO3-HF-HClO4, and then various heavy metal elements were tested. Specifically, As was determined by hydride generation-atomic fluorescence spectrometry (AFS-9800 atomic fluorescence spectrophotometer); Hg was determined by vapor generation-cold atomic fluorescence spectrometry (XGY-1011A atomic fluorescence spectrophotometer); Zn was determined by powder compression-X-ray fluorescence spectrometry (X-ray fluorescence spectrometer); and Cd, Cu, Ni, and Pb were determined by inductively coupled plasma mass spectrometry (iCAP QC inductively coupled plasma mass spectrometer).

[0049] The overall process of the mine ecological environment monitoring method in this embodiment is as follows: Figure 1 As shown, it includes the following steps:

[0050] Step 1: Calculate the potential ecological risk index (RI) based on the soil heavy metal content of sample points in the mining sample area, and use the multispectral remote sensing data of the sample points and the RI to perform regression fitting to obtain the RI inversion model.

[0051] Mining activities in polymetallic mining areas often lead to comprehensive pollution from multiple elements. Monitoring soil pollution in mining areas must comprehensively consider the potential risks to human health from various pollutants. This embodiment uses the Potential Ecological Risk Index (RI) soil pollution monitoring model to calculate the potential ecological risks of various heavy metal pollutants. The principle is to weight and sum the content of each element based on its degree of harm to human health and background values. The calculation formula is:

[0052] ,

[0053] In the formula, It is the first Potential ecological risk index for each sample point It is the first The first sample point Potential ecological risks of heavy metals; It is the first The toxicity response coefficient of a heavy metal reflects its toxicity to the human body and its sensitivity to environmental migration. It is the first At the nth sample point The content of various heavy metals; It is the first At the nth sample point Reference values ​​for the upper limit of pollution of several heavy metals. Toxicity response coefficient in this embodiment. Reference values ​​for upper limits of contamination based on Hakanson's test results. The Class III soil standard comes from the Soil Environmental Quality Standard (GB 15618-1995) issued by the Ministry of Ecology and Environment of China, which is the critical value for ensuring the growth of various plants.

[0054] Based on the spatial location of the sample points, spatial sampling was performed on six original bands of multispectral remote sensing data from Image2024: visible light (blue, green, red), near-infrared, and short-infrared (short-infrared band 1 and short-infrared band 2). These bands were used as independent variables, with a spectral range of 0.45–2.29 μm. Then, using the natural logarithm of the potential ecological risk index (RI) of the sample points as the dependent variable, a random forest method was used to regress the independent and dependent variables, resulting in an RI inversion model. Using this RI inversion model, the RI distribution for the entire region can be output based on the multispectral remote sensing data with the aforementioned six original bands as independent variables. This distribution is denoted as RI-ML.

[0055] In soil pollution research, data distortion can occur due to severely polluted areas. This embodiment uses natural logarithmic transformation as the dependent variable to maintain data quality. Furthermore, when actually using the RI inversion model, the output can be restored to RI by performing an exponential operation to obtain RI-ML.

[0056] Step 2: Perform band calculations on the multispectral remote sensing data of the mine sample area, and based on the calculated bands and the RI-ML obtained by the original bands through the RI inversion model, screen and construct the optimal band combination.

[0057] Step 2.1, Band Calculation. Using six original bands (red, green, blue, near-infrared, short-infrared band 1, and short-infrared band 2) with a resolution of 30m from Image2024 as independent variables, logarithmic and open-ended calculations were performed to obtain one-dimensional results. Additionally, based on the band calculation rules proposed by Wang et al., two-dimensional and three-dimensional band calculations were performed to obtain multi-dimensional results. The calculation methods and the number of bands are shown in Table 2, resulting in a total of 738 candidate bands.

[0058]

[0059] Step 2.2, Band Selection. Using the 738 candidate bands obtained from band calculations as independent variables and the RI-ML obtained from the original bands through the RI inversion model as the dependent variable, a random sampling method was used to calculate the correlation and multicollinearity of each band and the RI distribution. The top 50 bands with the highest correlation and multicollinearity (<5) were selected.

[0060] Step 2.3, band combination.

[0061] In step 2.2, the 50 bands obtained are arranged and combined according to the pre-selected quantity; then, each combination is used as a set of independent variables, and the potential ecological risk index RI is used as the dependent variable to perform linear regression. The optimal band combination is selected based on the root mean square error RSME of the linear regression model corresponding to each set of independent variables.

[0062] The optimal band combination obtained in this embodiment includes: , , , They are respectively:

[0063]

[0064]

[0065]

[0066] .

[0067] Step 3: Use the optimal band combination data of the mine sample area and RI-ML to perform regression fitting to obtain the RRI inversion model.

[0068] This embodiment uses the least partial squares regression method and performs regression fitting using band combination data of the mine sample area and RI-ML to obtain the RRI inversion model:

[0069] .

[0070] Step 4: Using the RRI inversion model and based on the optimal combination of remote sensing bands in the mine monitoring area, the remote sensing ecological risk index (RRI) is obtained by inverting the mine monitoring area.

[0071] By inputting the optimal band combination data of the mining area covered by satellites at different times into the RRI inversion model, the remote sensing ecological risk index (RRI) of the corresponding area can be obtained, reflecting the spatial characteristics of heavy metal pollution in the corresponding area.

[0072] Step 5: Calculate the greenness NDVI, humidity WET, and dryness NDBSI using multispectral remote sensing data of the mine monitoring area. Then, use principal component analysis to calculate RRI, NDVI, WET, and NDBSI, and use the first principal component as the pollution enhancement remote sensing ecological index PE-RSEI. In this embodiment, the cumulative contribution rate of the eigenvalues ​​of the first principal component PC1 is required to be ≥70%.

[0073] RSEI is a commonly used ecological assessment indicator, characterized by its simplicity, rich information, and strong time-bound applicability. However, for mining areas, especially non-ferrous metal mining areas, soil pollution is a crucial standard for environmental assessment, a factor that RSEI does not consider. This invention, based on the RSEI framework, replaces LST in the original RSEI with RRI obtained in step 4 to construct a new pollution-enhanced RSEI, denoted as PE-RSEI, with the following formula:

[0074]

[0075] Among them, PE-RSEI is the pollution-enhancing RSEI, and its value can directly characterize the current ecological environment quality of the area. NDVI is the Normalized Difference Vegetation Index, WET is the humidity index obtained by tasseling transformation, and NDBSI is the dryness index synthesized from the Normalized Building Index and the Normalized Bare Soil Index. The calculation formulas for the above three indicators are as follows:

[0076]

[0077]

[0078]

[0079]

[0080]

[0081]

[0082] In this embodiment, after calculating the remote sensing ecological risk index RRI, greenness NDVI, humidity WET, and dryness NDBSI in the aforementioned steps, the data of each index are first normalized, and then principal component analysis is used to calculate the principal components of the normalized RRI, NDVI, WET, and NDBSI.

[0083] Step 6: Conduct spatiotemporal analysis of the ecological environment of the mine monitoring area based on PE-RSEI.

[0084] When comparing and monitoring the ecological environment of a mining area based on the Pollution Enhancement Ecological Index (PE-RSEI) of different periods, the PE-RSEI of different periods can be further normalized to conduct time-series comparative analysis and spatial comparative analysis of the ecological environment of the mining area: the larger the PE-RSEI of a certain period, the higher the ecological environment quality of that period; the larger the PE-RSEI of a certain location, the higher the ecological environment quality of that location.

[0085] Therefore, the PE-RSEI for the corresponding year can be obtained from the multi-period multispectral remote sensing data obtained in step 1 using the method in step 3. Then, based on the historical spatial distribution and temporal changes of PE-RSEI, the spatiotemporal dynamic monitoring of the ecological environment of the mining area can be realized.

[0086] Example 2

[0087] This embodiment provides a mine ecological environment monitoring system that integrates soil heavy metal ecological indicators, including:

[0088] The RI calculation module is used to calculate the corresponding potential ecological risk index (RI) based on the soil heavy metal content of sample points in the mining sample area.

[0089] The RI fitting module is used to: perform regression fitting between multispectral remote sensing data of sample points and RI to obtain an RI inversion model;

[0090] The band calculation and filtering module is used to: perform band calculations on multispectral remote sensing data of the mine sample area, and filter and construct the best band combination based on the calculated bands and the RI-ML obtained by the RI inversion model of the original bands.

[0091] The RRI fitting module is used to obtain an RRI inversion model by performing regression fitting using the best band combination data of the mine sample area and RI-ML.

[0092] The RRI inversion module is used to: use the RRI inversion model and based on the optimal combination of remote sensing bands in the mining monitoring area to invert the remote sensing ecological risk index (RRI) of the mining monitoring area.

[0093] The principal component analysis module is used to: calculate the greenness NDVI, humidity WET, and dryness NDBSI of the mine monitoring area using multispectral remote sensing data, and then use principal component analysis to calculate RRI, NDVI, WET, and NDBSI, and extract the first principal component as the pollution enhancement remote sensing ecological index PE-RSEI.

[0094] The monitoring module is used to: perform spatiotemporal comparison of the ecological environment of the mine monitoring area based on PE-RSEI.

[0095] Example 3

[0096] This embodiment compares the method of Example 1 with existing monitoring methods based on experiments. Based on field test data and satellite imagery, the spatial distribution of the potential ecological risk index (RI) for the study area in 2024 was constructed using Kriging interpolation, machine learning inversion, and least partial squares regression methods, respectively. Figure 2 As shown. The spatial distribution of RI obtained using Kriging interpolation is as follows. Figure 2As shown in (a), high values ​​are mainly concentrated in the built-up area in the northwest and the industrial sites in the central and southern parts of the study area. The spatial distribution of RI in the study area was predicted using machine learning methods (RI-ML), as shown below. Figure 2 As shown in (b), the R² is 0.693 and the RMSE is 0.824, achieving good prediction accuracy. The high-value areas of RI-ML are mainly scattered industrial sites, and the distribution boundaries of these high-value areas closely approximate the surface land cover. The spatial distribution of RI (RRI) in the study area was predicted using least partial squares regression as follows: Figure 2 As shown in (c), the R² is 0.711 and the RMSE is 0.537, achieving good prediction accuracy.

[0097] The spatial distribution of PE-RSEI for each year from 1997 to 2024 is shown in [reference needed]. Figure 3 .Depend on Figure 3 Low-value areas of PE-RSEI are mainly concentrated in the built-up area in the northeast and scattered industrial sites in the south of the study area, exhibiting a clear road-line distribution pattern. High-value areas of PE-RSEI are mainly concentrated in the natural vegetation areas in the central and southern parts of the study area. From 1997 to 2024, the distribution characteristics of low and high values ​​in the study area did not change significantly, except for a large number of anomalous low-value areas appearing in the south in 2000, which is related to the quality of Landsat satellite data. The extremely low values ​​of PE-RSEI within the study area gradually shifted from the built-up area in the north to the mining areas in the central and southern parts of the study area, allowing the identification of the extent and change process of the mining areas from the spatial distribution of PE-RSEI. After 2000, the low-value areas of PE-RSEI in the north showed a clear characteristic of expansion over time, while the low-value areas of PE-RSEI in the south showed a clear characteristic of concentration within the various mining areas.

[0098] The above embodiments are preferred embodiments of this application. Those skilled in the art can make various changes or improvements based on them. Without departing from the overall concept of this application, these changes or improvements should fall within the scope of protection claimed in this application.

Claims

1. A method for monitoring the ecological environment of mines that integrates soil heavy metal ecological indicators, characterized in that, include: The potential ecological risk index (RI) is calculated based on the soil heavy metal content of sample points in the mining area, and the RI inversion model is obtained by regression fitting using multispectral remote sensing data of the sample points and RI. Band calculations were performed on the multispectral remote sensing data of the mining sample area, and the optimal band combination was selected and constructed based on the calculated bands and the RI-ML obtained by the RI inversion model of the original bands. The RRI inversion model is obtained by using the optimal band combination data of the mine sample area and RI-ML regression fitting. The Remote Sensing Ecological Risk Index (RRI) is obtained by using the RRI inversion model and based on the optimal combination of remote sensing bands in the mine monitoring area. The greenness NDVI, humidity WET, and dryness NDBSI of the mining monitoring area were calculated using multispectral remote sensing data. Then, principal component analysis was used to calculate RRI, NDVI, WET, and NDBSI, and the first principal component was used as the pollution enhancement remote sensing ecological index PE-RSEI. Spatiotemporal comparison of the ecological environment in the mine monitoring area was conducted based on PE-RSEI.

2. The method for monitoring the ecological environment of a mine according to claim 1, characterized in that, The calculation method for the potential ecological risk index RI is as follows: ， ； In the formula, It is the first Potential ecological risk index for each sample point It is the first The first sample point Potential ecological risks of heavy metals; It is the first The toxicity response coefficient of a heavy metal reflects its toxicity to the human body and its sensitivity to environmental migration. It is the first At the nth sample point The content of various heavy metals; It is the first At the nth sample point Reference values ​​for the upper limit of pollution of each heavy metal; n is the amount of heavy metal.

3. The method for monitoring the ecological environment of a mine according to claim 1, characterized in that, Six original bands of data from multispectral remote sensing data, including visible light, near-infrared, and short-infrared bands, were selected as independent variables, and the natural logarithm of the potential ecological risk index (RI) was used as the dependent variable. The random forest method was used to perform regression fitting on the independent and dependent variables to obtain the RI inversion model.

4. The method for monitoring the ecological environment of a mine according to claim 1, characterized in that, Band operations, filtering, and construction of optimal band combinations for multispectral remote sensing data include: Six original bands of data from multispectral remote sensing data were selected, including visible light, near infrared, and short infrared bands. One, two, and three bands were randomly selected from these bands to perform one-dimensional, two-dimensional, and three-dimensional operations, resulting in a total of 738 candidate band data. Using each candidate band data as the independent variable and the RI-ML obtained from the original band via the RI inversion model as the dependent variable, the correlation between the independent and dependent variables and the multicollinearity between the independent and dependent variables are calculated using a random sampling method. N bands with multicollinearity less than the preset value and the highest correlation with the dependent variable are selected from all candidate bands. The selected N bands are arranged and combined according to the pre-selected quantity; then, each combination is used as a set of independent variables, and RI-ML is used as the dependent variable for linear regression. The optimal band combination is selected based on the root mean square error (RSME) of the linear regression model corresponding to each set of independent variables.

5. The method for monitoring the ecological environment of a mine according to claim 4, characterized in that, The least partial squares regression method was used, and the band combination data of the sample points and RI were used for regression fitting to obtain the corresponding linear regression model.

6. The method for monitoring the ecological environment of a mine according to claim 4, characterized in that, The optimal band combinations include: , , , ;in, The values ​​represent the reflectance of the red band, green band, blue band, near-infrared band, short-infrared 1 band, and short-infrared band 2 bands obtained from the Landsat satellite multispectral sensor. , , , These represent the functions for square root calculation, natural logarithm calculation, predefined scheme calculation, and ratio calculation, respectively, and are expressed as follows: ； ； ； 。 7. The method for monitoring the ecological environment of a mine according to claim 1, characterized in that, Before using principal component analysis to calculate RRI, NDVI, WET, and NDBSI, the remote sensing ecological risk index RRI, greenness NDVI, humidity WET, and dryness NDBSI are calculated. The data of each index are then normalized, and principal component analysis is used to calculate the principal components of the normalized RRI, NDVI, WET, and NDBSI.

8. The method for monitoring the ecological environment of a mine according to claim 1, characterized in that, The spatiotemporal analysis of the ecological environment of the mine monitoring area based on PE-RSEI includes: the larger the PE-RSEI of a certain period, the higher the ecological environment quality of that period; the larger the PE-RSEI of a certain location, the higher the ecological environment quality of that location.

9. A mine ecological environment monitoring system integrating soil heavy metal ecological indicators, characterized in that, include: The RI calculation module is used to calculate the corresponding potential ecological risk index (RI) based on the soil heavy metal content of sample points in the mining sample area. The RI fitting module is used to: perform regression fitting between multispectral remote sensing data of sample points and RI to obtain an RI inversion model; The band calculation and filtering module is used to: perform band calculations on multispectral remote sensing data of the mine sample area, and filter and construct the best band combination based on the calculated bands and the RI-ML obtained by the RI inversion model of the original bands. The RRI fitting module is used to obtain an RRI inversion model by performing regression fitting using the best band combination data of the mine sample area and RI-ML. The RRI inversion module is used to: use the RRI inversion model and based on the optimal combination of remote sensing bands in the mining monitoring area to invert the remote sensing ecological risk index (RRI) of the mining monitoring area. The principal component analysis module is used to: calculate the greenness NDVI, humidity WET, and dryness NDBSI of the mine monitoring area using multispectral remote sensing data, and then use principal component analysis to calculate RRI, NDVI, WET, and NDBSI, and extract the first principal component as the pollution enhancement remote sensing ecological index PE-RSEI. The monitoring module is used to: perform spatiotemporal comparison of the ecological environment of the mine monitoring area based on PE-RSEI.

Images