Automated Monitoring Method for Vegetation Changes in Open-Pit Mining Areas Based on Long-Time Series Remote Sensing Data

Abstract

This invention provides an automated monitoring method for vegetation change in open-pit mining areas based on long-term time-series remote sensing data, belonging to the field of remote sensing and geographic information technology. The method includes the following steps: S1. Data acquisition and preprocessing: Acquire all data from the mining area during the study period, and obtain NDVI time-series data through cropping, NDVI calculation, maximum value synthesis, and filtering; S2. Template creation: Based on statistical analysis of the filtered NDVI time-series data, obtain the NDVI values ​​of vegetation and bare soil in the mining area, and generate NDVI templates for different vegetation change trajectory types by combining typical vegetation change patterns in open-pit mining areas; S3. Vegetation change trajectory type classification: Using the NDVI templates as training data, obtain the vegetation change trajectory type for each pixel in the mining area using a DTW-based KNN classification method; This invention can achieve a highly automated monitoring method for vegetation change in open-pit mining areas without sample training or parameter setting.

Description

Technical Field

[0001] This invention relates to the field of remote sensing and geographic information technology, specifically to an automated monitoring method for vegetation changes in open-pit mining areas based on long-term time-series remote sensing data. Background Technology

[0002] Surface vegetation is an important component of terrestrial ecosystems and plays a vital role in regional ecological environment. The growth process of vegetation is influenced by the synergistic effects of multiple factors, and vegetation can also respond to changes in the surrounding environment in a short period of time.

[0003] The development of remote sensing technology has facilitated research into the driving factors of vegetation change. The Normalized Difference Vegetation Index (NDVI) is one of the most commonly used indicators, showing a close positive correlation with vegetation productivity and vegetation cover. Long-term NDVI data can dynamically reflect vegetation change processes, thereby enabling the analysis of the environmental impacts of vegetation due to various natural and human activities.

[0004] Vegetation provides us with oxygen and food, playing a vital role in the balance of ecosystems. Vegetation change is closely related to global change and has therefore attracted much attention. Studies suggest that climate and human disturbance are the two main drivers of vegetation change. Remote sensing imagery, with its advantages of wide coverage, good data consistency, and strong real-time performance, has been widely used in the evaluation of vegetation change attribution. In the dynamic monitoring of vegetation remote sensing in open-pit mines, commonly used methods can be divided into two main categories: (1) threshold-based methods; and (2) disturbance detection algorithm-based methods. Threshold-based methods require sample training to obtain the corresponding threshold; while disturbance detection algorithm-based methods require setting multiple parameters for the algorithm used (such as Landtrendr, Bfast, CCDC, etc.). Both sample training and parameter setting require human operation, which requires a high level of professional experience and will affect the transferability of the method to some extent. Summary of the Invention

[0005] In view of this, the present invention provides an automated monitoring method for vegetation changes in open-pit mining areas based on long-term time-series remote sensing data. This method can achieve the goal of monitoring vegetation changes in open-pit mining areas without sample training or parameter setting. After data preprocessing, no manual operation is required, which reduces the difficulty of human operation and achieves the goal of highly automated monitoring of vegetation changes in open-pit mining areas. Moreover, the method is simple and reliable, the principle is simple, it is easy to implement, the work efficiency is high, and it is highly transferable.

[0006] To address the aforementioned technical problems, this invention provides an automated monitoring method for vegetation changes in open-pit mining areas based on long-term time-series remote sensing data, comprising the following steps:

[0007] S1. Data Acquisition and Preprocessing: Acquire all data from the mining area during the research period, and obtain NDVI time series data through cropping, NDVI calculation, maximum value synthesis method and filtering;

[0008] S2. Create a template: Based on the statistical analysis of the filtered NDVI time-series data, the vegetation (N) in the mining area is obtained. v ) and bare soil (N s The NDVI value of the open-pit mining area is used to generate NDVI templates for different vegetation change trajectory types, combined with typical vegetation change patterns in the open-pit mining area.

[0009] S3. Vegetation change trajectory type classification: Using NDVI templates as training data, the vegetation change trajectory type of each pixel in the mining area is obtained by using the DTW-based KNN classification method, thereby realizing automated monitoring of vegetation changes in open-pit mining areas.

[0010] Furthermore, when obtaining all data for the mining area study period in S1, all raw data for the mining area study period are downloaded via USGS.

[0011] Furthermore, the raw data includes remote sensing data acquired by all Landsat TM / ETM+ / OLI sensors during the mining area study period.

[0012] Furthermore, in S1, the maximum synthesized NDVI for each year during the study period of the mining area is obtained using the maximum value synthesis method, thereby obtaining the interannual NDVI time series data during the study period of the mining area.

[0013] Furthermore, during the filtering process in S1, the BISE-WT filter is specifically used to denoise the interannual NDVI time series, thereby obtaining the denoised NDVI time series data.

[0014] Furthermore, in S2, when generating NDVI templates for different vegetation change trajectory types based on typical vegetation change patterns in open-pit mining areas, the NDVI templates for each type are generated through permutations and combinations according to the stages corresponding to the vegetation change trajectory types and using the following relationships.

[0015] f = (ab)e -t / 2 +b

[0016] In the formula, t represents the year of recovery, and a = N s b = N v or N v ', N v N represents the NDVI value of the vegetation in the mining area. s The NDVI value of bare soil, N s and N v These represent 5% and 95% of all NDVI time series values ​​in the mining area, arranged in ascending order, respectively. v' is 0.8*N v f is the NDVI value of vegetation restoration.

[0017] Based on typical vegetation change patterns in open-pit mining areas, as well as the mining time and vegetation recovery status, mining areas are divided into the following five types:

[0018] PRP type: The stage is the vegetation restoration stage, characterized by continuous vegetation; mining is carried out before the study period, and the vegetation has recovered before the study period; NDVI remains at a high level throughout the study period.

[0019] PU type: The stage is the mining / topsoil backfilling stage, characterized by a consistently non-vegetated environment; mining occurred before the study period and no vegetation restoration was observed; NDVI remained at a low level throughout the study period.

[0020] DU type: The stage is from pre-mining to mining / topsoil backfilling stage, characterized by the change from vegetation to non-vegetation; mining was carried out during the study period, and no vegetation recovery was observed after mining. NDVI was at a high level in the early stage, followed by a sudden drop, after which NDVI was at a low level.

[0021] DR type: The stage is from pre-mining to vegetation restoration, characterized by the transformation from vegetation to non-vegetation and then back to vegetation; mining was carried out during the study period, and vegetation restoration was observed after mining. NDVI was at a high level in the early stage, followed by a sudden drop, after which NDVI gradually increased.

[0022] PRD type: The stage is from mining / topsoil backfilling to vegetation restoration, characterized by the transformation from non-vegetation to vegetation; mining was carried out before the study period, and vegetation restoration was observed. The NDVI was at a low level in the early stage, and then the NDVI gradually increased.

[0023] Specifically, each type of NDVI template is generated through permutations and combinations, as follows:

[0024] (1) PRP NDVI template: Generate values ​​of N respectively v and N v Two NDVI templates;

[0025] (2) PU's NDVI template: Generate a value of N s NDVI template;

[0026] (3) DU's NDVI templates: Generate templates with mining time at 25%, 50%, and 75% of the study period, resulting in a total of 6 templates according to the permutation and combination;

[0027] (4) NDVI templates for DR: Generate templates with mining time at 25%, 50%, and 75% of the study period, and vegetation restoration starts at 50% of the remaining time after mining. A total of 12 templates are generated according to the permutation and combination.

[0028] (5) PRD NDVI templates: Generate templates with vegetation restoration start time of 25%, 50%, and 75% of the study period, resulting in a total of 6 templates according to the permutation and combination.

[0029] Furthermore, in S3, the DTW-based KNN classification method includes the following steps:

[0030] S31. Use the NDVI template generated in S2 as sample data for sampling area classification;

[0031] S32. Calculate the DTW distance between the NDVI timing sequence and the NDVI template for each pixel in the sampling area;

[0032] S33. Classify the pixels as vegetation change trajectory types of the nearest DTW distance sample.

[0033] The beneficial effects of the above-mentioned technical solution of the present invention are as follows:

[0034] 1. The first step of the method of this invention is data acquisition and preprocessing. All Landsat surface reflectance images of the mining area during the research period are downloaded from the United States Geological Survey (USGS). Remote sensing imagery has advantages such as wide coverage, good data consistency, and strong real-time performance. Subsequently, NDVI time-series data is obtained through cropping, NDVI calculation, maximum value synthesis, and filtering. That is, the processed and corrected long-term remote sensing data can more fully reflect the spatiotemporal evolution of vegetation. Next, based on the statistical analysis of the filtered NDVI time-series data, the vegetation (N) of the mining area is obtained. v ) and bare soil (N s The NDVI value of the open-pit mining area is used to generate NDVI templates for different vegetation change trajectory types, combined with typical vegetation change patterns in the open-pit mining area. A method that does not require sample training or parameter setting is adopted, which requires no manual operation after data preprocessing, reducing the difficulty of human operation. Finally, using the NDVI templates as training data, the vegetation change trajectory type of each pixel in the mining area is obtained by using the KNN classification method based on DTW, thereby realizing automated monitoring of vegetation changes in the open-pit mining area. The method is simple, reliable, easy to implement, and efficient.

[0035] This invention utilizes all Landsat data from the open-pit mining area during the research period and employs a method that requires no sample training or parameter setting to achieve highly automated monitoring of vegetation changes in the open-pit mining area. After data preprocessing, no manual operation is required, reducing the difficulty of human operation.

[0036] 2. The present invention has strong portability and is applicable to automatic monitoring in fields such as urbanization, returning farmland to forest, abandonment of farmland, landscaping, and vegetation changes caused by logging and mining. Attached Figure Description

[0037] Figure 1 This is a typical NDVI variation pattern diagram of vegetation in open-pit mining areas in this invention;

[0038] Figure 2 This is a time-series example diagram of NDVI during the vegetation restoration phase in an embodiment of the present invention;

[0039] Figure 3 This is an example diagram of NDVI timing during the mining / topsoil backfilling stage in an embodiment of the present invention;

[0040] Figure 4 This is an example diagram of the NDVI time sequence from pre-mining to mining / topsoil backfilling stage in an embodiment of the present invention;

[0041] Figure 5 This is a time-series example diagram of NDVI from pre-harvest to vegetation restoration stage in an embodiment of the present invention;

[0042] Figure 6 This is a time-series NDVI example diagram from the mining / topsoil backfilling to vegetation restoration stage in an embodiment of the present invention;

[0043] Figure 7 This is an example diagram of the NDVI template for PRP in an embodiment of the present invention;

[0044] Figure 8 This is an example diagram of the NDVI template of the PU in an embodiment of the present invention;

[0045] Figure 9 This is an example diagram of the NDVI template for DR in an embodiment of the present invention;

[0046] Figure 10 This is an example diagram of the NDVI template of DU in an embodiment of the present invention;

[0047] Figure 11 This is an example diagram of the NDVI template of the PRD in this embodiment of the invention;

[0048] Figure 12 This is a classification diagram of vegetation change trajectory types in the final study area in this embodiment of the invention;

[0049] Figure 13 This is a flowchart of an embodiment of the present invention. Detailed Implementation

[0050] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the following will be described in conjunction with the accompanying drawings of the embodiments of the present invention. Figure 1-13 The technical solutions of the embodiments of the present invention will be clearly and completely described herein. Obviously, the described embodiments are only a part of the embodiments of the present invention, not all of them. All other embodiments obtained by those skilled in the art based on the described embodiments of the present invention are within the scope of protection of the present invention.

[0051] like Figure 1-13 As shown: An automated monitoring method for vegetation change in open-pit mining areas based on long-term remote sensing data includes the following steps:

[0052] S1. Data Acquisition and Preprocessing: Acquire all data from the mining area during the research period, and obtain NDVI time series data through cropping, NDVI calculation, maximum value synthesis method and filtering;

[0053] In S1, when obtaining all data for the mining area study period, all raw data for the mining area study period are downloaded via USGS.

[0054] The raw data includes remote sensing data acquired by all Landsat TM / ETM+ / OLI sensors during the mining area study period.

[0055] In S1, the maximum synthesized NDVI for each year during the study period of the mining area is obtained using the maximum value synthesis method, thereby obtaining the interannual NDVI time series data during the study period of the mining area.

[0056] In S1, the filtering process specifically uses a BISE-WT filter to denoise the interannual NDVI time series, thereby obtaining the denoised NDVI time series data.

[0057] S2. Create a template: Based on the statistical analysis of the filtered NDVI time-series data, the vegetation (N) in the mining area is obtained. v ) and bare soil (N s The NDVI value of the open-pit mining area is used to generate NDVI templates for different vegetation change trajectory types, combined with typical vegetation change patterns in the open-pit mining area.

[0058] Among them, typical vegetation change patterns in open-pit mining areas are as follows: Figure 1 As shown, according to Figure 1 The vegetation change process in open-pit mining areas can be divided into four stages: pre-mining, vegetation restoration, mining / topsoil backfilling, and vegetation restoration.

[0059] Based on the NDVI variation patterns of open-pit mines, as well as the mining time and vegetation recovery status, the mines are divided into the following five types, as shown in Table 1:

[0060]

[0061] Table 1

[0062] In S2, when generating NDVI templates for different vegetation change trajectory types based on typical vegetation change patterns in open-pit mining areas, the NDVI templates for each type are generated through permutations and combinations according to the stages corresponding to the vegetation change trajectory types and using the following relationships.

[0063] f = (ab)e -t / 2 +b

[0064] In the formula, t represents the year of recovery, and a = N s b = N v or N v ', N v N represents the NDVI value of the vegetation in the mining area. s The NDVI value of bare soil, N s and N v These represent 5% and 95% of all NDVI time series values ​​in the mining area, arranged in ascending order, respectively. v ' is 0.8*N v f is the NDVI value of vegetation restoration.

[0065] Table 2 below shows the NDVI values ​​for each stage more clearly.

[0066]

[0067] Based on the stages corresponding to each vegetation change trajectory type in Table 1, we plan to use the NDVI values ​​corresponding to each stage in Table 2 to generate NDVI templates for each type through permutation and combination. Specifically:

[0068] (1) PRP NDVI template: Generate values ​​of N respectively v and N v Two NDVI templates;

[0069] (2) PU's NDVI template: Generate a value of N s NDVI template;

[0070] (3) DU's NDVI templates: Generate templates with mining time at 25%, 50%, and 75% of the study period, resulting in a total of 6 templates according to the permutation and combination;

[0071] (4) NDVI templates for DR: Generate templates with mining time at 25%, 50%, and 75% of the study period, and vegetation restoration starts at 50% of the remaining time after mining. A total of 12 templates are generated according to the permutation and combination.

[0072] (5) PRD NDVI templates: Generate templates with vegetation restoration start time of 25%, 50%, and 75% of the study period, resulting in a total of 6 templates according to the permutation and combination.

[0073] S3. Vegetation change trajectory type classification: Using NDVI templates as training data, the vegetation change trajectory type of each pixel in the mining area is obtained by using the DTW-based KNN classification method, thereby realizing automated monitoring of vegetation changes in open-pit mining areas.

[0074] In S3, the classification method based on DTW and KNN includes the following steps:

[0075] S31. Use the NDVI template generated in S2 as sample data for sampling area classification;

[0076] S32. Calculate the DTW distance between the NDVI timing sequence and the NDVI template for each pixel in the sampling area;

[0077] S33. Classify the pixels as vegetation change trajectory types of the nearest DTW distance sample.

[0078] The specific study area is the open-pit mining area in Jindui Town, Shaanxi Province. All raw data from 2000 to 2020 for the mining area were downloaded from USGS, namely remote sensing data acquired by all Landsat TM / ETM+ / OLI sensors.

[0079] Following the methods described in S2 above, the NDVI template for the mining area was generated, as follows: Figure 7 As shown, the generated NDVI template is a sample image of PRP, that is, it generates values ​​of N respectively. v and N v Two NDVI templates; such as Figure 8 As shown, the generated NDVI template is a sample image of PU, that is, it generates a value of N. s NDVI templates; such as Figure 9 As shown, the generated templates are NDVI examples for DR, showing templates generated at 25%, 50%, and 75% of the mining period, with vegetation restoration starting at 50% of the remaining time after mining. A total of 12 templates are generated based on these combinations. Figure 10 The image shows a sample NDVI template for DU, generating templates for mining times at 25%, 50%, and 75% of the study period. A total of six templates are generated based on these combinations. Specifically, when the mining time is at 25%, 50%, and 75% of the study period, the generated value is N.v The three NDVI templates, and the values ​​generated when the mining time was 25%, 50%, and 75% of the study period, were N respectively. v Three NDVI templates; such as Figure 11 As shown, the generated NDVI template sample diagram of PRD is a template that generates six templates with vegetation restoration starting time of 25%, 50%, and 75% of the study period.

[0080] Subsequently, the vegetation change trajectory type classification of the study area was obtained according to method S3, specifically as follows: Figure 12 As shown.

[0081] Based on the aforementioned automated monitoring method for vegetation change in open-pit mining areas using long-term time-series remote sensing data, a classification of vegetation change trajectory types in the open-pit mining areas is ultimately generated. Following the above process, a highly automated monitoring of vegetation change in open-pit mining areas can be achieved, requiring no manual operation after data preprocessing, thus reducing the difficulty of human intervention. Taking the open-pit mining area of ​​Jindui Town, Shaanxi Province as an example, the final classification of vegetation change trajectory types in the study area is as follows: Figure 12 As shown.

[0082] The first step of the method of this invention is data acquisition and preprocessing. All Landsat surface reflectance images of the mining area during the research period are downloaded from the United States Geological Survey (USGS). Remote sensing imagery has advantages such as wide coverage, good data consistency, and strong real-time performance. Subsequently, NDVI time-series data is obtained through cropping, NDVI calculation, maximum value synthesis, and filtering. That is, the processed and corrected long-term remote sensing data can more fully reflect the spatiotemporal evolution of vegetation. Next, based on the statistical analysis of the filtered NDVI time-series data, the vegetation (N) of the mining area is obtained. v ) and bare soil (N s The NDVI value of the open-pit mining area is used to generate NDVI templates for different vegetation change trajectory types, combined with typical vegetation change patterns in the open-pit mining area. A method that does not require sample training or parameter setting is adopted, which requires no manual operation after data preprocessing, reducing the difficulty of human operation. Finally, using the NDVI templates as training data, the vegetation change trajectory type of each pixel in the mining area is obtained by using the KNN classification method based on DTW, thereby realizing automated monitoring of vegetation changes in the open-pit mining area. The method is simple, reliable, easy to implement, and efficient.

[0083] This invention utilizes all Landsat data from the open-pit mining area during the research period and employs a method that requires no sample training or parameter setting to achieve highly automated monitoring of vegetation changes in the open-pit mining area. After data preprocessing, no manual operation is required, reducing the difficulty of human operation. The method is simple and reliable.

[0084] Furthermore, the method of the present invention is highly portable and applicable to automatic monitoring in fields such as urbanization, reforestation, farmland abandonment, landscaping, and vegetation changes caused by logging and mining.

[0085] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. An automated monitoring method for vegetation change in open-pit mining areas based on long-term time-series remote sensing data, characterized in that, Includes the following steps: S1. Data Acquisition and Preprocessing: Acquire all data from the mining area during the research period, and obtain NDVI time series data through cropping, NDVI calculation, maximum value synthesis method and filtering; S2. Create templates: Based on the statistical analysis of the NDVI time-series data, obtain the NDVI values ​​of vegetation Nv and bare soil Ns in the mining area, and combine them with the typical vegetation change patterns in the open-pit mining area to generate NDVI templates for different vegetation change trajectory types. S3. Vegetation change trajectory type classification: Using NDVI templates as training data, the vegetation change trajectory type of each pixel in the mining area is obtained by using the KNN classification method based on DTW, thereby realizing automated monitoring of vegetation changes in open-pit mining areas. In S2, when generating NDVI templates for different vegetation change trajectory types based on typical vegetation change patterns in open-pit mining areas, the NDVI templates for each type are generated through permutations and combinations according to the stages corresponding to the vegetation change trajectory types and using the following relationships. In the formula, t represents the year of restoration, a = Ns, b = Nv or Nv', Nv is the NDVI value of vegetation in the mining area, Ns is the NDVI value of bare soil, Ns and Nv are 5% and 95% of the time-series NDVI values ​​of all pixels in the mining area in ascending order, respectively, Nv' is 0.8*Nv, and f is the NDVI value of vegetation restoration.

2. The automated monitoring method for vegetation change in open-pit mining areas based on long-term remote sensing data according to claim 1, characterized in that, When obtaining all data for the mining area study period in S1, all raw data for the mining area study period are downloaded via USGS.

3. The automated monitoring method for vegetation change in open-pit mining areas based on long-term remote sensing data according to claim 2, characterized in that, The raw data includes remote sensing data acquired by all Landsat™ / ETM+ / OL I sensors during the mining area study period.

4. The automated monitoring method for vegetation change in open-pit mining areas based on long-term time-series remote sensing data according to claim 3, characterized in that, In S1, the maximum synthesized NDVI for each year during the study period of the mining area is obtained using the maximum value synthesis method, thereby obtaining the interannual NDVI time series data during the study period of the mining area.

5. The automated monitoring method for vegetation change in open-pit mining areas based on long-term time-series remote sensing data according to claim 4, characterized in that, During the filtering process in S1, the BI SE-WT filter is used to denoise the interannual NDVI time series, thereby obtaining the denoised NDVI time series data.

6. The automated monitoring method for vegetation change in open-pit mining areas based on long-term time-series remote sensing data according to claim 1, characterized in that, In S3, the classification method based on DTW and KNN includes the following steps: S31. Use the NDVI template generated in S2 as sample data for sampling area classification; S32. Calculate the DTW distance between the NDVI time sequence and the NDVI template for each pixel in the sampling area; S33. Classify the pixels as vegetation change trajectory types of the nearest DTW distance sample.

Images