DEM data-based polder area boundary extraction method

Abstract

A polder area boundary extraction method based on DEM data comprises the following steps: A1, performing depression filling processing on input projection DEM data, eliminating local low-depression noise, and improving data quality; a2, setting an elevation threshold value by calculating an elevation mean value and a standard deviation of the grids after depression filling, and extracting a region higher than the threshold value to generate a preliminary binary grid map; and A3, converting the raster data into vector polygons, and calculating the area of each polygon. And A4, extracting the polygon with the maximum area as final polder area output, and storing the polygon as an element class. According to the method, DEM elevation statistical characteristics and a GIS spatial analysis technology are combined, finally, through automatic model construction, precise polder area extraction, connectivity optimization and small area block intelligent fusion are achieved, multi-area data can be efficiently processed in batches, and data integrity and reliability are improved.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of information technology, and can be used for extracting polder boundaries in plain water network areas. Specifically, the present application relates to an automatic polder boundary extraction method. BACKGROUND

[0002] A polder refers to an area surrounded by water on all sides and having a lower elevation, which can only be used for production and life by relying on the protection of dikes. These areas are usually represented as low-lying areas surrounded by higher-elevation dikes in DEM (Digital Elevation Model) data. Accurate identification and boundary extraction of polders are of great significance to water management, agricultural planning, flood control, and ecological monitoring. However, it is difficult to effectively deal with complex terrain changes and data noise.

[0003] Although various GIS and remote sensing methods can be used for polder extraction, traditional polder extraction methods rely on manual interpretation or simple elevation threshold screening, but still face many challenges, such as the discontinuity or fracture of polder dikes due to image data errors or DEM precision problems, or discontinuity caused by actual conditions, resulting in incomplete boundary extraction, which makes it difficult to effectively deal with complex terrain changes and data noise.

[0004] Therefore, in order to solve the above problems, the present application proposes an efficient and automated polder boundary extraction method based on DEM data and combined with GIS analysis. SUMMARY

[0005] The present application provides a polder boundary extraction method based on DEM data, which uses DEM data, combines GIS spatial analysis techniques and statistical principles, and constructs an automated polder extraction and optimization model.

[0006] The present application adopts the following technical solutions: A polder boundary extraction method based on DEM data, comprising the following steps: A1: First, fill the low-lying areas of the input projected DEM data to eliminate local low-lying noise and improve data quality; A2: Then, calculate the mean and standard deviation of the elevation of the filled low-lying area grid, set an elevation threshold, and extract the area above the threshold to generate a preliminary binary grid. To improve connectivity, the model further expands the extracted area and assigns a unique ID to the connected areas using the region grouping tool; A3: Next, convert the grid data to vector polygons and calculate the area of each polygon. To remove small discrete polygons, the model sets 10% of the total area as a screening threshold to screen and merge polygons with an area smaller than the threshold; A4: Finally, the maximum area polygon is extracted as the final polder output and saved as a feature class.

[0007] The entire process is based on ModelBuilder and can be applied in batch mode to complete the automatic extraction and screening of multiple feature classes, ensuring that the extracted polders have strong connectivity, reasonable area, and complete data. The algorithm can adapt to various DEM data to improve the degree of automation and reduce manual intervention.

[0008] The step A1 specifically comprises the following steps: 1. Input data: DEM data containing polders, coordinate system is the projected coordinate system; output data is stored in a file GDB format database (such as: DATA.GDB); 2. The following steps use Modelbuilder to build a model, add input data variables, name: DEM, which is the projected DEM data; 3. DEM fill preprocessing, add Fill tool, input raster data: DEM, output raster data: DEM_FILL; The step A2 specifically comprises the following steps: 4. Use the standard deviation method to set the threshold according to the statistical properties of the elevation, the threshold setting method is: mean plus 1 times the standard deviation (Mean + StdDev). Add 2 Get Raster Properties tools to extract Mean (average value obtained from Get Raster Properties) and StdDev (standard deviation obtained from Get Raster Properties 2) of DEM. Input raster data: DEM_FILL; output raster data: Dem_Mean, Dem_StdDev respectively; 5. Add Calculate Value tool. Use Calculate Value tool to add Mean and StdDev to get the threshold, expression is: %Dem_Mean% + %Dem_StdDev%, output is a numeric variable, variable name is: Threshold; 6、Add Con Conditional Function tool, based on the condition expression to extract the pixels in the raster that meet the condition, the expression is: value > %Threshold%, Input true raster or constant value is set to 0, Input false raster or constant value is set to 1, the input raster data is: DEM_FILL, the output raster data is: DEM_EXTRACT, the output raster data is a binary graph of 0, 1; 7、Add Expand tool, which enhances the connectivity of the target class by expanding the pixels of the specified class in the raster, and processes the polder area into a continuous closed region. The input raster data is: DEM_Extract, the output raster data is: DEM_Expand, and the Number of Cells parameter is set to 1; 8、Add Region Group tool, which can group the extracted regions according to spatial connectivity (i.e. pixels connected to the same group), and assign a unique ID value to each group. The input raster data is: DEM_Expand, the output raster data is: DEM_Region, the Number of neighbors to use parameter is set to 4, and the Zone grouping method parameter is set to within; 9、Add Raster to Polygon tool, which converts raster data to vector data, the input raster data is: DEM_Region, and the output vector data is: Dem_Polygon; The step A3 specifically includes the following steps: 10、Add Summary Statistics tool, which summarizes the area of the output polygon layer, selects the field Shape_Area to be counted, selects Sum as the statistical type, inputs the vector data: Dem_Polygon, and outputs the table: Area_Sum; 11、Add Get Field Value tool to read the value of the SUM_Shape_Area field in the Area_Sum table, input table: Area_Sum, field selection: SUM_Shape_Area, data type selection: Double; The output variable name is: value 12、Add Calculate Value tool, set the threshold of 10% of the total area, the expression is: %value% / 10; The output variable name is: Area_threshold; 13. Add Make Feature Layer tool, convert the feature class to a layer for easy operation, input feature class: Dem_Polygon, output layer: Dem_Polygon_Layer; 14. Add Select Layer By Attribute tool, create a new selection set, input layer: Dem_Polygon_Layer, expression: Shape_Area <= %Area_threshold%, output layer: Small_Polygons_Select; 15. Add Eliminate tool, based on the new selection set Small_Polygons_Select layer created in the last step, merge the selected features into adjacent polygons, input layer: Small_Polygons_Select, output feature class: Polygon_Eliminate; The step A4 specifically includes the following steps: 16. Add Summary Statistics tool, summarize the output Polygon_Eliminate feature class, select the field Shape_Area to be counted, select Max for the statistics type, input feature class data: Polygon_Eliminate, output table: Max_Area; 17. Add Join Field tool, input table: Polygon_Eliminate, input join field: Shape_Area; join table: Max_Area, join field: MAX_Shape_Area; output layer variable name: Polygon_Eliminate_Join; 18. Add Make Feature Layer tool, input layer: Polygon_Eliminate_Join, output layer: Polygon_Eliminate_Layer; 19. Add Select Layer By Attribute tool, create a new selection set, input layer: Polygon_Eliminate_Layer, expression: MAX_Shape_Area IS NOT NULL, output layer: Max_Polygon_Select; 20. Add Copy Features tool, export the selected polder polygon as a feature class. Input: Max_Polygon_Select, output feature class: Polder 21. Add Get Raster Properties tool, extract the DEM raster size (CELLSIZE), select the attribute type parameter as CELLSIZEX or CELLSIZEY, and the output variable name is: CELLSIZE; 22. Add Buffer tool to restore the extracted polder area reduced by one pixel due to the use of Expand tool. Input vector data: Polder, output vector data: Polder_Buffer, distance from CELLSIZE.

[0009] 23. Select other feature classes that need to be processed to run the model for processing, or use the feature iterator to loop the above steps 2-22 to process the data, and at the same time, some links can be set as parameters for user application.

[0010] The method has the advantages that the method combines the DEM elevation statistical characteristics with the GIS spatial analysis technology, especially uses the elevation mean and standard deviation to set the threshold value, expands the outer boundary through the Expand tool for connection, and finally restores the extracted boundary range through the Buffer tool, so that the accurate extraction of the polder area, the connectivity optimization and the intelligent fusion of small blocks are realized through the automatic model construction, multi-region data can be efficiently and batch processed, and the data integrity and reliability are improved. BRIEF DESCRIPTION OF DRAWINGS

[0011] Figure 1 For the original DEM data Figure 2 For the binary graph (dem_extract) Figure 3 For the Dem_Expand graph Figure 4 For the Dem_Polygon Figure 5 For the Polygon_Eliminate graph after removing the fine polygon Figure 6 For the generated polder area graph (Polder_Buffer) Figures 7-9 For the model schematic diagram divided into three parts. DETAILED DESCRIPTION

[0012] The application will be described in detail below in combination with specific embodiments.

[0013] The method of the application uses DEM data, combines GIS spatial analysis technology and statistical principles, and constructs an automatic polder extraction and optimization model, including the following steps A1-A4: A1: Refer to Figure 7,First, the input projection DEM data is processed to ,eliminate local low-lying noise and improve data quality; Specifically includes steps 1-3: 1. Input data: Contains DEM data of the polder area ( Figure 1 ), the coordinate system is the projection coordinate system; the output data is stored in the database in the GDB format file (such as: DATA.GDB); 2. The following steps use Modelbuilder to build the model and add an input data variable named DEM, which is the projected DEM data. 3. DEM filling preprocessing, add Fill tool, input raster data: DEM, output raster data: DEM_FILL; A2: Reference Figure 7 ,Then, by calculating the elevation mean and standard deviation of the grid after filling, setting the elevation threshold, and extracting the areas above the threshold, a preliminary binary raster map is generated;,To improve connectivity, the model further expands the extracted areas and assigns unique IDs to the connected areas through the,region grouping tool; Specifically including steps 4-9: 4. Use the standard deviation method to set a threshold based on the statistical characteristics of the elevation. The threshold setting method is: mean plus 1 times the standard deviation (Mean + StdDev). Add two Get Raster Properties tools to extract the DEM's Mean (the mean obtained from Get Raster Properties) and StdDev (the standard deviation obtained from Get Raster Properties 2). The input raster data is: DEM_FILL; the output raster data are: Dem_Mean and Dem_StdDev. 5. Add the Calculate Value tool. Use the Calculate Value tool to add Mean and StdDev to get the threshold value. The expression is: %Dem_Mean% + %Dem_StdDev%. The output is a numeric variable named: Threshold. 6. Add the Con conditional function tool to extract pixels that meet the conditions in the raster based on the conditional expression: value > %Threshold%, Input true raster or constant value is set to 0, Input false raster or constant value is set to 1, input raster data is: DEM_FILL, output raster data is: DEM_EXTRACT ( Figure 2), the output raster data is a binary image with 0 and 1; 7. Add the Expand tool, which enhances the connectivity of the target class by expanding the pixels of the specified class in the raster, and processes the patches into continuous closed regions. The input raster data is DEM_Extract, and the output raster data is DEM_Expand ( Figure 3 ), and the Number of Cells parameter is set to 1; 8. Add the Region Group tool, which can group the extracted regions according to spatial connectivity (i.e., pixels that are connected are grouped into the same group), and assign a unique ID value to each group. The input raster data is DEM_Expand, and the output raster data is DEM_Region, the Number of neighbors to use parameter is set to 4, and the Zone grouping method parameter is set to within; 9. Refer to Figure 8 , add the Raster to Polygon tool to convert the raster data to vector data, the input raster data is DEM_Region, and the output vector data is Dem_Polygon ( Figure 4 ); A3: Then, convert the raster data to vector polygons and calculate the area of each polygon. To remove small discrete polygons, set 10% of the total area as the screening threshold, and screen and merge the polygons with an area smaller than the threshold; Specifically, steps 10-15 are included: 10. Add the Summary Statistics tool to summarize the area of the output polygon layer, select the field Shape_Area to be counted, and select Sum as the statistical type, the input vector data is Dem_Polygon, and the output table is Area_Sum; 11. Add the Get Field Value tool to read the SUM_Shape_Area field value in the Area_Sum table, the input table is Area_Sum, the field selection is SUM_Shape_Area, and the data type selection is Double; the output variable name is value; 12. Add the Calculate Value tool, set 10% of the total area as the threshold, and the expression is %value% / 10; the output variable name is Area_threshold; 13, Add Make Feature Layer tool, convert the feature class to layer for convenient operation, input feature class: Dem_Polygon, output layer: Dem_Polygon_Layer; 14, Add Select Layer By Attribute tool, create a new selection set, input layer: Dem_Polygon_Layer, expression: Shape_Area <= %Area_threshold%, output layer: Small_Polygons_Select; 15, Add Eliminate tool, based on the Small_Polygons_Select layer created in the last step, merge the selected features into adjacent polygons, input layer: Small_Polygons_Select, output feature class: Polygon_Eliminate Figure 5 ); A4: Finally, the largest area polygon is extracted as the final output of the polder, and saved as a feature class. The entire process is based on ModelBuilder, which can be applied in batch processing mode to complete the automatic extraction and selection of multiple feature classes, ensuring the connectivity of the extracted polder, reasonable area and complete data. The algorithm can adapt to various DEM data to improve the degree of automation and reduce manual intervention.

[0014] Specifically, steps 16-23 are included: 16, Add Summary Statistics tool, summarize the output Polygon_Eliminate feature class, select the field Shape_Area to be counted, select Max for the statistics type, input feature class data: Polygon_Eliminate, output table: Max_Area; 17, refer to Figure 9 , add Join Field tool, input table: Polygon_Eliminate, input join field: Shape_Area; Join table: Max_Area, join field: MAX_Shape_Area; Output layer variable name: Polygon_Eliminate_Join; 18, Add Make Feature Layer tool, input layer: Polygon_Eliminate_Join, output layer: Polygon_Eliminate_Layer; 19. Add the Select Layer By Attribute tool and create a new selection set. The input layer is: Polygon_Eliminate_Layer, the expression is: MAX_Shape_Area IS NOT NULL, and the output layer is: Max_Polygon_Select; 20. Add the Copy Features tool to export the selected polder polygons as a feature class. Input: Max_Polygon_Select, Output feature class: Polder; 21. Add the Get Raster Properties tool to extract the cell size (CELLSIZE) of the DEM. Select CELLSIZEX or CELLSIZEY for the attribute type parameter. The output variable name is: CELLSIZE. 22. Added Buffer tool to recover the problem of shrinking the extracted polder area by one pixel due to using Expand tool. Input vector data: Polder, output vector data: Polder_Buffer ( Figure 6 ), the distance comes from CELLSIZE; Select other feature classes that need to be processed and run the model for processing, or use the feature iterator to loop the above steps to process the data. At the same time, some links can be set as parameters to facilitate user application.

[0015] The automated polder extraction model constructed by this invention significantly improves efficiency by at least 10 to 15 times compared to manual operations. The entire process is automated, reducing human error. It is suitable for large-scale data processing and can process multiple study areas in a loop.

[0016] It should be understood that those skilled in the art can make improvements or changes based on the above description, and all such improvements and changes should fall within the scope of protection of the appended claims of the present invention.

Claims

1. A method for extracting polder area boundaries based on DEM data, characterized in that: The following steps are involved: A1: Perform depression filling processing on the input projected DEM data to eliminate local low-lying noise and improve data quality; A2: By calculating the mean and standard deviation of the grid after filling, setting an elevation threshold, and extracting areas above the threshold, a preliminary binary raster map is generated. To improve connectivity, the model further expands the extracted areas and assigns unique IDs to connected areas using the region grouping tool. A3: Convert raster data into vector polygons and calculate the area of ​​each polygon. To remove discrete polygons with too small an area, the model sets a screening threshold of 10% of the total area and filters and merges polygons with an area smaller than this threshold. A4: Extract the polygon with the largest area as the final polder output and save it as a feature class.

2. The method for extracting polder area boundaries based on DEM data according to claim 1, characterized in that: The step A1 specifically includes the following steps: A1-1. Input data: Contains DEM data of the polder area, with a projected coordinate system; output data is stored in a database in GDB format; A1-2. Use ModelBuilder to build the model and add an input data variable named DEM, which is the projected DEM data. A1-3. DEM filling preprocessing, add the Fill tool, the input raster data is: DEM, the output raster data is: DEM_FILL.

3. The method for extracting polder area boundaries based on DEM data according to claim 1, characterized in that: The step A2 specifically includes the following steps: A2-4. Use the standard deviation method to set the threshold based on the statistical characteristics of the elevation. Add two Get Raster Properties tools to extract the Mean and StdDev of the DEM respectively. The input raster data is: DEM_FILL; the output raster data are: Dem_Mean and Dem_StdDev. A2-5. Add the Calculate Value tool. Use the Calculate Value tool to add Mean and StdDev to get the threshold value. The expression is: %Dem_Mean% + %Dem_StdDev%. The output is a numeric variable named Threshold. A2-6. Add the Con conditional function tool to extract pixels that meet the conditions in the raster based on the conditional expression: value > %Threshold%. Set the Input true raster or constant value to 0 and the Input false raster or constant value to 1. The input raster data is: DEM_FILL, and the output raster data is: DEM_EXTRACT. The output raster data is a binary image of 0 and 1. A2-7. Add the Expand tool. This tool expands the pixels of a specified class in the raster to enhance the connectivity of the target class and process the polder area into a continuous closed area. The input raster data is: DEM_Extract, the output raster data is: DEM_Expand, and the Number of Cells parameter is set to 1. A2-8. Add the Region Group tool, which groups the extracted regions according to spatial connectivity, assigning connected pixels to the same group and assigning a unique ID value to each group. The input raster data is DEM_Expand, and the output raster data is DEM_Region. Set the Number of neighbors to use parameter to 4 and the Zone grouping method parameter to within. A2-9. Add the Raster to Polygon tool to convert raster data into vector data. The input raster data is: DEM_Region, and the output vector data is: Dem_Polygon.

4. The method for extracting polder area boundaries based on DEM data according to claim 1, characterized in that: The step A3 specifically includes the following steps: A3-10. Add the Summary Statistics tool to summarize the area of ​​the output polygon layer. Select the Shape_Area field to be counted, select Sum for the statistic type, input vector data: Dem_Polygon, and output table: Area_Sum. A3-11. Add the Get Field Value tool to read the value of the SUM_Shape_Area field in the Area_Sum table. Enter the Area_Sum table, select the SUM_Shape_Area field, and select the Double data type. Name the output variable value. A3-12. Add the Calculate Value tool and set the threshold to 10% of the total area using the expression: %value% / 10. The output variable is named: Area_threshold; A3-13. Add the Make Feature Layer tool. To facilitate calculations, convert the feature class into a layer. The input feature class is Dem_Polygon, and the output layer is Dem_Polygon_Layer. A3-14. Add the Select Layer By Attribute tool and create a new selection set. The input layer is Dem_Polygon_Layer, the expression is Shape_Area <= %Area_threshold%, and the output layer is Small_Polygons_Select. A3-15. Add the Eliminate tool. Create a new selection set, Small_Polygons_Select, based on the previous step. Merge the selected features into adjacent polygons. The input layer is Small_Polygons_Select, and the output feature class is Polygon_Eliminate.

5. The method for extracting polder area boundaries based on DEM data according to claim 1, characterized in that: The step A4 specifically includes the following steps: A4-16. Add the Summary Statistics tool to summarize the output Polygon_Eliminate feature class. Select the Shape_Area field to be counted, select Max for the statistic type, input the Polygon_Eliminate feature class data, and output the Max_Area table. A4-17. Add the Join Field tool. Input table: Polygon_Eliminate, input join field: Shape_Area; join table: Max_Area, join field: MAX_Shape_Area; output layer variable name: Polygon_Eliminate_Join; A4-18. Add the Make Feature Layer tool, with the input layer being Polygon_Eliminate_Join and the output layer being Polygon_Eliminate_Layer. A4-19. Add the Select Layer By Attribute tool and create a new selection set. The input layer is Polygon_Eliminate_Layer, the expression is MAX_Shape_Area IS NOT NULL, and the output layer is Max_Polygon_Select. A4-20. Add the Copy Features tool to export the selected polder polygons as a feature class. Input: Max_Polygon_Select, Output feature class: Polder A4-21. Add the Get Raster Properties tool to extract the DEM grid size CELLSIZE. Select CELLSIZEX or CELLSIZEY for the attribute type parameter. The output variable name is: CELLSIZE. A4-22. Added the Buffer tool to recover from the problem of the extracted polder area being reduced by one pixel due to the use of the Expand tool. Input vector data: Polder, output vector data: Polder_Buffer, distance is derived from CELLSIZE.

6. The method for extracting polder area boundaries based on DEM data according to claim 3, characterized in that: In step A2-4, the threshold is set by adding 1 times the standard deviation to the mean value.

Images