A method for identifying soil erosion patches based on spatial data of soil erosion intensity

Abstract

This invention discloses a method for identifying soil erosion patches based on spatial data of soil erosion intensity. It enables the inheritance and connection of spatial boundaries between different scale levels of soil erosion intensity data, accurately locating areas where soil erosion occurs. This allows for targeted guidance in implementing soil erosion control measures according to land type, soil erosion intensity level, and specific needs. It meets the application requirements of soil erosion monitoring results from macro to micro scales, providing strong support for soil and water conservation work such as the assessment of soil and water conservation rate targets, the layout and effectiveness evaluation of comprehensive soil erosion control projects, and has significant application prospects and socio-economic benefits.

Description

Technical Field

[0001] This invention belongs to the field of soil and water conservation and remote sensing monitoring technology, and relates to a method for identifying soil erosion patches based on spatial data of soil erosion intensity. Background Technology

[0002] Current remote sensing monitoring results for soil erosion suffer from unclear erosion intensity boundaries and a lack of coherence among various data sets. This makes it difficult to locate soil erosion patches across different land use types, posing challenges to the precise decomposition and implementation of soil erosion control tasks and the evaluation of the effectiveness of precise soil and water conservation measures. Therefore, establishing a method for "landing" spatial data on soil erosion to achieve accurate identification of soil erosion patches is crucial. This would enable targeted guidance for implementing soil erosion control measures based on land type, erosion intensity level, and specific needs. It would also meet the application requirements of soil erosion monitoring results from macro to micro scales, providing strong support for soil and water conservation work, including performance evaluation of soil and water conservation targets, the layout of key control projects, and effectiveness assessment. This approach offers significant socio-economic and environmental benefits. Summary of the Invention

[0003] The purpose of this invention is to provide a method for identifying soil erosion patches based on spatial data of soil erosion intensity, which solves the problems of unclear erosion intensity boundaries and difficulty in locating soil erosion patches of different land use types in existing remote sensing monitoring results of soil erosion.

[0004] The technical solution adopted in this invention is a method for identifying soil erosion patches based on spatial data of soil erosion intensity. The specific operation steps are as follows:

[0005] Step 1: According to the "Classification and Grading Standard for Soil Erosion", the soil erosion modulus A raster data is divided into five soil erosion intensity levels: mild (500≤A<2500), moderate (2500≤A<5000), severe (5000≤A<8000), extremely severe (8000≤A<15000), and extremely severe (A≥15000). The soil erosion area under each intensity level is obtained by statistically analyzing the soil erosion data of the study area.

[0006] Step 2: Convert the land use vector data into land use raster data, overlay the land use raster data and the soil erosion modulus A raster data from Step 1, and statistically obtain the soil erosion area of ​​different land use types under each intensity level.

[0007] Step 3: Based on the soil erosion intensity level in Step 1, establish a new soil erosion intensity level raster layer; divide the new soil erosion intensity according to the land use vector data to obtain the first layer of soil erosion vector pattern for each new soil erosion intensity level;

[0008] The new soil erosion intensity is divided into three levels: mild I: 500≤A<2500, moderate II: 2500≤A<8000, and severe or above III: A≥8000;

[0009] Step 4: Based on Step 3, and combined with the multi-scale segmentation method, the first layer of soil erosion map patches of the new soil erosion intensity level is further segmented to obtain the second layer of soil erosion vector map patches of the new soil erosion intensity level. The second layer of soil erosion vector map patches includes: mild I soil erosion map patches, moderate II soil erosion map patches, and severe or above III soil erosion map patches.

[0010] Step 5: Overlay the soil erosion modulus A raster data from Step 1 onto the second layer of soil erosion vector patches, calculate the average soil erosion modulus value inside each soil erosion vector patch, and obtain soil erosion vector patches with average erosion modulus information.

[0011] Step 6: Serially number the land parcels in the land use vector data so that each land use parcel has a parcel number DKBH. Convert the parcel number DKBH into a raster layer and overlay the soil erosion vector plot with average erosion modulus information obtained in Step 5. Obtain the raster data value of the parcel number DKBH corresponding to each soil erosion vector plot.

[0012] Spatially connect the soil erosion vector patch and the land use vector data using the same plot number DKBH raster data value to obtain the land use type information of each soil erosion patch, thus obtaining a soil erosion vector patch with land use type information.

[0013] Step 7: Based on the average soil erosion modulus value in Step 5, the soil erosion vector map patches obtained in Step 6 are further divided according to the "Soil Erosion Classification and Grading Standard" (SL 190-2007) to obtain soil erosion map patches under 5 soil erosion intensity levels.

[0014] Step 8: Verify whether the average erosion modulus values ​​of the soil erosion patches of each intensity level obtained in Step 7 are within the range of the corresponding intensity level values ​​in the "Classification and Grading Standard for Soil Erosion" (SL 190-2007). If the accuracy of the soil erosion patches of each intensity level is higher than 95%, it indicates that the spatial boundaries based on the soil erosion intensity level raster data are well inherited and connected in space. If the accuracy is lower than 95%, it is necessary to return to Step 4, adjust the segmentation scale, and correct the second layer of soil erosion vector map patches.

[0015] Step 9: Calculate the soil erosion area of ​​the soil erosion patches under the five soil erosion intensity levels obtained in Step 7, and compare the relative error with the soil erosion area obtained in Step 1 and Step 2; if the relative error is generally less than 5%, it meets the accuracy requirements; if the relative error is greater than 5%, return to Step 4 to correct the second layer of patch segmentation results.

[0016] Step 10: After meeting the spatial location accuracy and area accuracy requirements of Steps 8 and 9, merge all the soil erosion patches under the five soil erosion intensity levels obtained in Step 7 into a new layer to obtain soil erosion patches that conform to the "Soil Erosion Classification and Grading Standard" (SL 190-2007) and have land use type information.

[0017] Step 11: Further merge the soil erosion patches obtained in Step 10 according to the conditions of the same land use type and adjacent location, and eliminate those smaller than 400m according to the quality control requirements of the "2021 Technical Guidelines for Dynamic Monitoring of Soil Erosion". 2 The following small patches ultimately yielded the results of soil erosion patch identification under different land use types.

[0018] The invention is further characterized in that,

[0019] Step 3 is as follows:

[0020] Multi-scale segmentation refers to segmenting image pixels based on the spectral heterogeneity between adjacent pixels and a set spectral heterogeneity threshold, forming a target object composed of multiple homogeneous pixels, thereby achieving the classification of the target object. This method comprehensively considers the spectral characteristics of ground objects and the spatial, textural, and other geometric features and structural information of image objects.

[0021] The land use vector data and the new soil erosion intensity level raster layer from step 3 are overlaid separately. Using the eCognition object-oriented classification tool, the new soil erosion intensity is segmented according to the land use boundary using a multi-scale segmentation method. The segmentation scale value must be greater than the product of the horizontal and vertical pixels of the soil erosion intensity level raster layer to obtain the first layer of soil erosion vector pattern for each new soil erosion intensity level.

[0022] Step 4 is as follows:

[0023] Using the eCognition object-oriented classification tool, and after testing at different cutting scales, the optimal cutting scales (thresholds) for the multi-scale segmentation method in the raster layers of three new soil erosion intensity levels (mild I, moderate II, and severe III) were found to be 40, 200, and 200, respectively.

[0024] Step 6 is as follows:

[0025] The DKBH raster layer and the soil erosion patches obtained in step 5 are overlaid. The tabular display method in ArcGIS is then used to statistically analyze the zoning. This method involves obtaining the DKBH raster data value corresponding to each soil erosion patch and displaying the results in a table. Using the layer connection method in ArcGIS, the soil erosion patches and land use vector data are connected according to the same DKBH patch value to obtain the land use type information for each soil erosion patch.

[0026] Step 7 is as follows:

[0027] Based on the average soil erosion modulus value from step 5, the soil erosion vector plots obtained in step 6 are further divided according to the "Classification and Grading Standard for Soil Erosion" (SL 190-2007). Specifically: soil erosion plots in the Slight I layer remain Slight; soil erosion plots in the Moderate II layer are divided into Moderate and Severe soil erosion plots based on an average erosion modulus value of 5000; soil erosion plots in the Strong and Above III layer are divided into Extremely Severe and Extremely Severe soil erosion plots based on an average erosion modulus value of 8000. Finally, soil erosion plots under five soil erosion intensity levels are obtained.

[0028] Step 8 is as follows:

[0029] According to the soil erosion modulus range of the five intensity levels in the "Classification and Grading Standard for Soil Erosion" (SL 190—2007), verify whether the average erosion modulus value of the soil erosion patches reclassified in step 7 is within the standard grading range. Count the number of erroneous soil erosion patches in the light, moderate, strong, very strong, and severe levels respectively. If the accuracy rate of the reclassified soil erosion patches for each intensity level is higher than 95%, the requirements are met; if the accuracy rate is lower than 95%, return to step four to adjust the segmentation scale and correct the second layer of soil erosion vector map patches.

[0030] Step 11 is as follows:

[0031] The soil erosion patches obtained in step 10 are merged according to the conditions of the same land use type and adjacent location to obtain merged soil erosion patches. The area (m²) of each soil erosion patch is calculated. 2 ), filter by area attribute for areas smaller than 400m² 2 The following small patches, smaller than 400m, can be removed using the Eliminate tool in ArcGIS. 2 The smaller polygons below are merged into the adjacent polygons with the largest area or the longest shared boundary.

[0032] During the elimination of small soil erosion patches, a small number of patches could not be eliminated due to their spatially discrete distribution. For these few patches, manual removal was performed, and the relative error between the total soil erosion area of ​​the removed patches and the total soil erosion area from step 1 was evaluated. Verification showed that these spatially discrete patches with an area less than 400 m² were suitable for elimination. 2 The following soil erosion patches are extremely small and will not affect the overall accuracy of soil erosion patch identification. Ultimately, patches smaller than 400m will be eliminated. 2 The following small patches are used to obtain the soil erosion patch identification results under different land use types.

[0033] This invention transforms remote sensing monitoring results of soil erosion into soil erosion control units that meet the needs of practical governance applications. It constructs a scale system conversion method of "county-soil erosion map patch basic unit", realizes the inheritance and connection of spatial boundaries between different management scale levels, accurately locates areas where soil erosion occurs, fills the gap in the application of soil erosion monitoring results from macro scale to micro-level fine scale, and provides important support for meeting the needs of soil and water conservation prevention and supervision, comprehensive management, soil and water conservation function evaluation, and soil and water conservation target responsibility assessment.

[0034] The beneficial effects of this invention are:

[0035] 1. By combining soil erosion raster data and land use vector data for hierarchical segmentation, the accuracy of soil erosion patch identification under different intensity levels of different land use types can be effectively improved, thereby improving the spatial accuracy of soil erosion patch identification among different land uses.

[0036] 2. Establishing a method for identifying soil erosion patches can realize the inheritance and connection of spatial boundaries between different management scale levels of the "soil erosion remote sensing monitoring - basic spatial unit for comprehensive soil erosion control", which is an innovation at the methodological level;

[0037] 3. It efficiently and easily transforms the spatial distribution results of remote sensing monitoring of soil erosion into the smallest management unit that can meet the deployment of comprehensive soil erosion control measures. It has low requirements for the working environment and location, has wide regional applicability, and reduces the difficulty of soil erosion patch identification. Attached Figure Description

[0038] Figure 1 This is a flowchart illustrating the method for identifying soil erosion patches based on spatial data of soil erosion intensity, as described in this invention.

[0039] Figure 2 This is a diagram illustrating the soil erosion intensity classification rules of the present invention.

[0040] Figure 3 This is an example diagram of the second layer of soil erosion vector pattern in this invention;

[0041] Figure 4 Example images of soil erosion patches at five intensity levels;

[0042] Figure 5 Example image of soil erosion patch identification results;

[0043] Figure 6(a) is a map showing the distribution of field verification points in Xingguan District according to Embodiment 7 of the present invention;

[0044] Figure 6(b) is a map showing the distribution of field verification points in Rongjiang County according to an embodiment of the present invention;

[0045] Figure 6(c) is a map showing the distribution of field verification points in Qinglong County according to an embodiment of the present invention. Detailed Implementation

[0046] The present invention will be further described in detail below with reference to the embodiments and accompanying drawings, but the embodiments of the present invention are not limited thereto.

[0047] Example 1

[0048] The soil erosion intensity spatial data-based method for identifying soil erosion patches proposed in this invention provides a new approach to solving the problem of how to transform the spatial distribution results of soil erosion monitored by remote sensing into the smallest management unit that can meet the requirements of comprehensive soil erosion control measures. This invention will provide important support for achieving precise deployment of control measures and meeting the needs of refined and efficient comprehensive soil erosion control.

[0049] See Figure 1 This invention is a method for identifying soil erosion patches using spatial data of soil erosion intensity. The specific steps of this method are as follows:

[0050] Step 1: Selection of Study Area

[0051] Guizhou Province is characterized by a large area of ​​soil erosion, high intensity, and scattered distribution. Its karst and rocky desertification areas exhibit distinct soil erosion features, resulting in a complex and multifaceted situation. Therefore, its soil erosion research and control are highly representative nationwide. Taking into account the following three principles: ① distribution in both karst and non-karst areas; ② severe and representative soil erosion; ③ distribution in both national-level key soil erosion prevention and control areas and provincial-level soil erosion monitoring areas, three pilot counties were selected to conduct soil erosion patch identification experiments. The accuracy of soil erosion patch identification was compared to verify the regional applicability of the method of this invention.

[0052] The pilot counties are: ① National key areas for soil and water conservation and karst areas: Qixingguan District of Bijie City and Qinglong County of Qianxinan Buyi and Miao Autonomous Prefecture, Guizhou Province; ② Provincial soil and water conservation monitoring areas and non-karst areas: Rongjiang County of Qiandongnan Miao and Dong Autonomous Prefecture.

[0053] Step 2: Extraction of Soil and Water Loss Area

[0054] Using raster data of soil erosion modulus A from three pilot counties, soil erosion intensity was classified into five levels according to the "Classification and Grading Standard of Soil Erosion": mild (500≤A<2500), moderate (2500≤A<5000), severe (5000≤A<8000), extremely severe (8000≤A<15000), and extremely severe (A≥15000). The soil erosion area under each intensity level in each pilot county was counted.

[0055] The land use vector data of each pilot county was converted into land use raster data. The land use raster data and soil erosion raster data of different intensity levels were overlaid to obtain the soil erosion area of ​​different land use types under each intensity level.

[0056] Step 3: Method for Identifying Soil Erosion Patches

[0057] In each pilot county, a new soil erosion intensity level raster layer was established based on the soil erosion intensity level in step 2; the new soil erosion intensity was segmented according to the land use vector data to obtain the first layer of soil erosion vector plots for each new soil erosion intensity level; the new soil erosion intensity was divided into three levels ( Figure 2 The severity levels are: Mild I: 500≤A<2500, Moderate II: 2500≤A<8000, and Severe or worse III: A≥8000.

[0058] By combining multi-scale segmentation methods, the first layer of soil erosion map patches of the new soil erosion intensity level is further segmented to obtain the second layer of soil erosion vector map patches. Figure 3 The second layer of soil erosion vector maps includes: mild soil erosion (Level I), moderate soil erosion (Level II), and severe soil erosion (Level III).

[0059] Step 4: Re-extraction of information from soil erosion patches

[0060] By overlaying the soil erosion modulus A raster data from step 2 onto the second layer of soil erosion vector patches, the average soil erosion modulus value inside each soil erosion vector patch is calculated to obtain soil erosion vector patches with average erosion modulus information.

[0061] Each pilot county's land use vector plots are sequentially numbered, giving each plot a unique identifier value named DKBH (plot number). The DKBH field is converted into a raster layer, and soil erosion vector plots are overlaid to obtain the plot number DKBH value corresponding to each soil erosion vector plot. The soil erosion vector plots and land use vector data are spatially connected using the same plot number DKBH value to obtain the land use type information for each soil erosion plot, resulting in a soil erosion vector plot with land use type information.

[0062] Finally, based on the average soil erosion modulus value in the soil erosion patches, the soil erosion vector plots of each new soil erosion intensity level raster layer were further divided according to the "Soil Erosion Classification and Grading Standard" (SL 190-2007) to obtain soil erosion patches under 5 soil erosion intensity levels in the pilot county.

[0063] Step 5: Evaluation of the accuracy of soil erosion patch identification

[0064] In terms of spatial location accuracy, the average erosion modulus values ​​of soil erosion patches obtained in step 4 were examined to see if they were within the range of the corresponding intensity level values ​​in the "Classification and Grading Standard for Soil Erosion" (SL 190-2007). The results showed the accuracy of soil erosion patches in the pilot county under the five soil erosion intensity levels, as shown in Table 1.

[0065] Table 1. Statistical table of accuracy of soil erosion maps.

[0066]

[0067] The accuracy rates of soil erosion maps in the three pilot counties were 98.27%, 97.33%, and 99.99%, respectively, with spatial location accuracy reaching over 95%, meeting the accuracy requirements and achieving a good and accurate conversion of spatial data from remote sensing monitoring of soil erosion.

[0068] In terms of area accuracy, the relative error between the soil erosion patches identified by this invention and the total soil erosion area reported by the remote sensing soil erosion monitoring grid is less than 2% at each intensity level, and the relative error of the total soil erosion area is less than 1% (Table 2). The relative errors of soil erosion area for different land use types and soil erosion area at each intensity level are controlled within 5% (Table 3), all meeting the accuracy requirements.

[0069] Table 2. Relative Accuracy of Identification Area of ​​Soil Erosion Patches at Different Intensity Levels

[0070]

[0071] Table 3. Relative Accuracy of Identified Area of ​​Soil Erosion Plots at Different Intensity Levels for Different Land Use Types

[0072]

[0073]

[0074] Step Six: Results of Soil Erosion Patches Identification

[0075] After all soil erosion patches in the pilot counties met the requirements for spatial location accuracy and area accuracy, the soil erosion patches under the five soil erosion intensity levels obtained in step 4 were merged into a new layer to obtain soil erosion patches that conform to the "Classification and Grading Standard for Soil Erosion" (SL 190-2007) and have land use type attribute information. Figure 4 ).

[0076] Further integration was carried out based on the conditions of similar land use type and adjacent location, and the areas smaller than 400m were eliminated in accordance with the quality control requirements of the "2021 Technical Guidelines for Dynamic Monitoring of Soil and Water Erosion". 2 The following are small patches. During the removal of these small patches, a small number could not be removed due to their spatially discrete distribution. For these few patches, manual removal was performed, and the relative error between the total soil erosion area of ​​the removed patches and the total soil erosion area in step 2 was evaluated (Table 4). Verification showed that the three pilot counties had spatially discrete patches with areas less than 400 m². 2 The following soil erosion patches are extremely small and will not affect the overall accuracy of soil erosion patch identification. Ultimately, patches smaller than 400m will be eliminated. 2 The following small patches are used to obtain the soil erosion patch identification results under different land use types in the pilot county. Figure 5 The final number of soil erosion patches in each pilot county is shown in Table 5.

[0077] Table 4. Relative accuracy of identified area for soil erosion patches of different intensity levels after minimization.

[0078]

[0079]

[0080] Table 5 Comparison of Soil Erosion Plot Identification Results with the Number of Original Land Use Vector Plots

[0081] County (city, district) Number of land parcels in land use vector data (number of parcels) Number of soil erosion patches identified (number) Qinglong County 29395 18051 Rongjiang County 83237 53124 Qixingguan District, Bijie City 61580 65342

[0082] Step 7: Field verification and validation of soil erosion patches

[0083] This invention conducts field verification of soil erosion identification results to validate the accuracy of soil erosion patch identification in pilot counties. Based on the "Technical Specification for Remote Sensing Monitoring of Soil and Water Conservation (SL592-2012)," 0.5% of the total number of soil erosion patches in each pilot county were verified in the field. In Qixingguan District, due to the scattered distribution of soil erosion patches in orchards and construction land, the area of ​​soil erosion was relatively small and all were of a mild erosion nature; therefore, cultivated land, forest land, and grassland, where soil erosion mainly occurs, were selected for field verification. In Rongjiang County and Qinglong County, cultivated land, forest land, grassland, orchards, and construction land were selected for field verification, respectively.

[0084] Finally, as shown in Figures 6(a)-(c), the number of field verification survey points in the three pilot counties were 337, 260, and 110, respectively; the verification accuracy was 91.96%, 91.54%, and 87.27% (Table 6), respectively. The accuracy of soil erosion patches was higher than 85%. Therefore, the soil erosion patch boundary obtained by the soil erosion patch identification method of this invention can provide strong support for soil and water conservation work such as the assessment of the responsibility system for soil and water conservation rate targets, the layout of key governance projects, and the evaluation of effectiveness. This will enable targeted guidance for implementing soil erosion control measures according to land type, grade, and needs.

[0085] Table 6. Statistical Table of Field Verification Accuracy of Soil Erosion Map Patch

[0086]

Claims

1. A method for identifying soil erosion map patches based on spatial data of soil erosion intensity, characterized in that, The specific operation steps are as follows: Step 1: According to the "soil erosion classification standard", the soil erosion modulus A grid data is divided into five soil erosion intensity levels, i.e. light 500≤A<2500, moderate 2500≤A<5000, strong 5000≤A<8000, very strong 8000≤A<15000 and severe A≥15000, and the soil erosion area under each intensity level is obtained by statistical analysis of the soil erosion data in the study area; Step 2: Convert the land use vector data to land use grid data, superimpose the land use grid data and the soil erosion modulus A grid data of step 1, and obtain the soil erosion area of different land use types under each intensity level by statistical analysis; Step 3: According to the soil erosion intensity level of step 1, a new soil erosion intensity level grid layer is established; according to the land use vector data, the new soil erosion intensity is divided into three levels: light I: 500≤A<2500, moderate II: 2500≤A<8000 and strong III: A≥8000; Step 4: According to step 3, combined with the multi-scale segmentation method, the first layer of soil and water loss vector map patches of the new soil erosion intensity level is further segmented to obtain the second layer of soil and water loss vector map patches of the new soil erosion intensity level, which includes light I soil and water loss map patches, moderate II soil and water loss map patches and strong III soil and water loss map patches; Step 5: Using the second layer of soil and water loss vector map patches, the soil erosion modulus A grid data of step 1 is superimposed to calculate the average soil erosion modulus value inside each soil and water loss vector map patch, and the soil and water loss vector map patch with average erosion modulus information is obtained; Step 6: The land blocks in the land use vector data are continuously numbered, so that each land use block has a block number DKBH, and the block number DKBH is converted into a grid layer, which is superimposed with the soil and water loss vector map patch with average erosion modulus information obtained in step 5 to obtain the grid data value of the block number DKBH corresponding to each soil and water loss vector map patch; The soil and water loss vector map patch and the land use vector data are spatially connected through the same block number DKBH grid data value to obtain the land use type information of each soil and water loss map patch, and the soil and water loss vector map patch with land use type information is obtained; Step 7: According to the average soil erosion modulus value of step 5, the soil and water loss vector map patch obtained in step 6 is divided again according to the "soil erosion classification standard" (SL 190-2007) to obtain the soil and water loss map patch under five soil erosion intensity levels. ​ Step 8: Check whether the average erosion modulus values of each intensity level of the soil erosion map patches obtained in step 7 are within the value range of the corresponding intensity level in the "Soil Erosion Classification and Grading Standard" (SL 190-2007). If the accuracy of each intensity level of the soil erosion map patches is higher than 95%, it indicates that the spatial inheritance and connection of the soil erosion intensity level raster data is better. If the accuracy is lower than 95%, return to step 4 to adjust the segmentation scale and correct the second layer of soil erosion vector map patches. Step 9: Calculate the soil erosion area of the soil erosion map patches under the five soil erosion intensity levels obtained in step 7, and compare the relative error with the soil erosion area obtained in steps 1 and 2. If the relative error is generally lower than 5%, it meets the accuracy requirement. If the relative error is higher than 5%, return to step 4 to correct the second layer of map patch segmentation results. Step 10: After meeting the spatial position accuracy and area accuracy requirements in steps 8 and 9, combine all the soil erosion map patches under the five soil erosion intensity levels obtained in step 7 into a new layer to obtain the soil erosion map patches that meet the "Soil Erosion Classification and Grading Standard" (SL 190-2007) and have land use type information. Step 11: The soil erosion map patches obtained in step 10 are further fused according to the same land use type and adjacent location conditions, and small patches less than 400 m 2 are eliminated according to the quality control requirements of the "2021 Annual Soil Erosion Dynamic Monitoring Technical Guidelines", and finally the soil erosion map patch identification results under different land use types are obtained.

2. The soil erosion intensity spatial data-based soil loss map patch identification method according to claim 1, characterized by, Step 3 is as follows: Overlay the land use vector data and the new soil erosion intensity level raster layer, and use the eCognition object-oriented classification tool to segment the new soil erosion intensity according to the land use boundary using the multi-scale segmentation method. The segmentation scale value should be greater than the product of the horizontal and vertical pixel numbers of the soil erosion intensity level raster layer. This will obtain the first layer of soil erosion vector map patches for each new soil erosion intensity level.

3. The soil erosion intensity spatial data-based soil loss map patch identification method according to claim 1, characterized by, Step 4 is as follows: Use the eCognition object-oriented classification tool to test different cutting scales and obtain the best segmentation scale threshold values for the multi-scale segmentation method in the light intensity I, medium intensity II, and intense III new soil erosion intensity level raster layers, which are 40, 200, and 200 respectively.

4. The soil erosion intensity spatial data-based soil loss map patch identification method according to claim 1, characterized by, Step 6 is as follows: Overlay the DKBH raster layer and the soil erosion map patches with average erosion modulus information obtained in step 5, and use the table display partition statistics method in the ArcGIS tool to obtain the raster data values of each soil erosion map patch corresponding to the plot number DKBH and display the results in table form. Use the layer connection method in the ArcGIS tool to connect the soil erosion map patches and the land use vector data according to the same raster data values of the plot number DKBH to obtain the land use type information of each soil erosion map patch.

5. The soil erosion intensity spatial data-based soil loss map patch identification method according to claim 1, characterized by, Step 8 is as follows: According to the soil erosion modulus value range of 5 intensity grades in the Standard for Classification and Grading of Soil Erosion (SL 190-2007), the average erosion modulus values of the water and soil loss map patches of each intensity grade reclassified in step 7 are tested to see whether they are within the standard grading value range, and the number of water and soil loss map patches of error in the light, moderate, strong, extremely strong and severe grades is respectively counted; if the accuracy rates of the water and soil loss map patches of each intensity grade reclassified are all higher than 95%, it is indicated that the requirements are met; if the accuracy rates are lower than 95%, step 4 is returned to, the segmentation scale is adjusted, and the second layer water and soil loss vector map patch is corrected. 6.The method of identifying a soil erosion map patch based on soil erosion intensity spatial data according to claim 1, wherein, Step 11 is specifically as follows: The soil erosion patches obtained in step 10 are merged according to the conditions of the same land use type and adjacent location to obtain merged soil erosion patches. The area of ​​each soil erosion patch is calculated, and those with an area of ​​less than 400m² are filtered out by area attribute. 2 Small patches smaller than 400m can be removed using the removal tools in ArcGIS. 2 Small patches are merged into adjacent patches with the largest area or the longest shared boundary; In the process of eliminating small patches, there are a small amount of patches that cannot be eliminated due to spatial dispersion; for these small patches, they are removed by human beings, and the relative error of the total soil erosion area of the water and soil loss map patches after removal and the total soil erosion area of step 1 is evaluated.

Images