Drainage basin remote sensing water body segmentation extraction method based on topographic information
Through the basin remote sensing water body segment extraction method based on topographic information, the basin division index and water body extraction index are constructed using digital elevation model and building land vector data, which solves the problem that complex terrain and artificial buildings shadows affect the accuracy of water body information extraction in traditional methods, and realizes the accuracy and timeliness of segmented water body extraction in the entire basin.
Patent Information
- Application Number
- CN202411707053.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-27
- Publication Date
- 2025-02-25
AI Technical Summary
Traditional water basin water resource distribution information acquisition is low in time and costly, and factors such as complex terrain and artificial buildings affect the accuracy of water body information extraction. A single water body extraction index is difficult to ensure the accuracy of water body extraction within the basin.
The basin remote sensing water body segment extraction method based on topographic information is adopted. Through digital elevation model data, building land vector data and optical remote sensing image data, the basin division index and water body extraction index are constructed, and the appropriate water body extraction algorithm is selected by region to realize the whole basin segmented water body extraction.
The accurate extraction of water bodies is achieved in complex environments, the timeliness and accuracy of obtaining water distribution information in the basin is improved, the cost is reduced, and it has important application value.
Smart Images
Figure CN119513222A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to a method for extracting water bodies by segmentation based on topographic information from remote sensing of a watershed, and belongs to the technical fields of remote sensing technology and hydrological monitoring. Background Art
[0002] The distribution information of water resources in the basin plays an indispensable role in many fields such as water resources management, environmental protection, disaster warning and social and economic sustainability, and is of great significance for promoting sustainable social and economic development and maintaining ecological and environmental security. Traditional basin water resources distribution information mainly relies on manual mapping and statistical water body information, which is time-sensitive, expensive and costly. With the large-scale application of remote sensing technology, the use of remote sensing data for water body extraction has also achieved remarkable results. For different water body information, a variety of water body extraction indexes have been studied, which greatly improves the accuracy of water body distribution information extraction and plays an important role in practical applications. However, due to the complex natural environment of water body distribution in the basin, the complex terrain faced in the process of water body extraction in the basin mainly includes: obvious elevation changes in the upper reaches of the basin, mountain shadows, tree occlusions, etc.; the middle and lower reaches of the river pass through urban areas or population gathering areas, and the shadows and occlusions of artificial buildings affect the extraction of water body information. It is difficult to guarantee the accuracy of water body extraction within the basin using a single water body extraction index, which restricts the accuracy of water body extraction in the basin. Summary of the invention
[0003] The purpose of the present invention is to overcome the above-mentioned shortcomings and provide a method for segmented extraction of water bodies in a river basin based on terrain information. The method comprehensively considers the differences in water body extraction under different terrain environments, and realizes accurate extraction of water bodies in complex environments by constructing remote sensing water body extraction methods for different areas of different river basin segments.
[0004] The technical solution adopted by the present invention is: A method for extracting water bodies by segmentation based on remote sensing of a watershed based on terrain information comprises the following steps: S1. Collect DEM (digital elevation model) data, building land vector data, and optical remote sensing image data of the target watershed; S2. Use ArcGIS software to calculate the slope of the collected DEM data to obtain the slope value S ; S3. Calculate the normalized digital elevation model value, and use the normalized digital elevation model value and slope value S to construct a watershed division index based on terrain changes WSDI ; S4. Divide the basin into three areas: upper, middle and lower reaches according to the value range of the basin division index; S5. The middle and lower reaches are further divided into built-up areas and non-built-up areas according to the built-up land vector data. The water body extraction area is divided into the upper reaches, the middle reaches built-up areas, the middle reaches non-built-up areas, the lower reaches built-up areas, and the lower reaches non-built-up areas in combination with the watershed division and the built-up land vector data; S6. Select different water extraction algorithms for different water extraction areas, and use the shadow water index in the upper reaches of the basin. SWI , the modified normalized difference water index was selected in the built-up areas in the middle and lower reaches of the basin MNDVI , select the normalized water index in the non-built-up areas in the middle and lower reaches of the basin NDVI ; S7. Comprehensively construct a comprehensive regional remote sensing water extraction method based on watershed division, building area division, and water index selection, and determine the segmented water extraction algorithm for the entire watershed: , In the formula, WSDI is the watershed division index, a , b The threshold for watershed division, C is a built-up area or a non-built-up area, n1, n2, n3, n4, and n5 are water extraction thresholds, and the threshold values are determined based on the experience of the area and the actual situation; S8. For the water bodies within the watershed, perform regional extraction according to the formula obtained in step S7.
[0005] In the above method, the calculation formulas for the normalized digital elevation model value and the watershed division index described in step S3 are as follows: , In the formula, ND_DEM is the normalized digital elevation model value, DEM is the digital elevation model value, DEM max is the maximum value of the digital elevation model within the watershed. DEM min It is the minimum value of the digital elevation model within the watershed; , In the formula, WSDI is the watershed division index, ND_DEM is the normalized digital elevation model value, S is the slope value.
[0006] The watershed division criteria described in step S4 are as follows: , In the formula, a , b is a constant and is related to the watershed in which it is located.
[0007] In step S5, the construction area and non-construction area are divided as follows: .
[0008] The calculation formulas for each water body index in step S6 are as follows: , , , In the formula, B is the reflectance value of the blue band, G is the reflectance value of the green band, NIR is the reflectance value in the near infrared band, WORLD is the reflectivity value in the mid-infrared band.
[0009] The final water body extraction algorithm in step S7 is: , In the formula, B is the reflectance value of the blue band, G is the reflectance value of the green band, NIR is the reflectance value in the near infrared band, WORLD is the reflectance value in the mid-infrared band, n1, n2, n3, n4, and n5 are the water extraction thresholds, and the threshold values are determined based on the empirical values of the area and the actual situation.
[0010] The present invention uses digital elevation model (DEM) data to extract information on slope and aspect, integrates multiple terrain parameters, assists other parameters, analyzes the watershed, forms a watershed division index, divides the watershed according to the range of different indexes, and further subdivides the area according to whether there are buildings around the water body in different divided areas. In each subdivided area, the best water body extraction index is selected to achieve accurate water body extraction in the entire watershed. The present invention is based on terrain information, comprehensively considers the differences in water body extraction under different terrain environments, and achieves accurate water body extraction under complex environments by constructing a regional remote sensing water body extraction method based on terrain information. It is of great significance for the comprehensive acquisition and utilization of water body distribution information over a large range, and has obvious exemplary application in other regions. BRIEF DESCRIPTION OF THE DRAWINGS
[0011] Figure 1 is a flow chart of the method of the present invention; Figure 2 The digital elevation model data obtained in the embodiment; Figure 3 The building land vector data obtained in the embodiment; Figure 4 The Sentinel-2 optical remote sensing image data obtained in the embodiment; Figure 5 is the result of slope calculation in the embodiment; Figure 6 It is the numerical calculation result of the normalized digital elevation model in the embodiment; Figure 7 The calculation result of the watershed division index in the embodiment; Figure 8 This is a diagram of the watershed division results in the embodiment; Figure 9 It is the distribution map of the building area in the watershed in the embodiment; Figure 10 This is the water body result of a river basin in Example. DETAILED DESCRIPTION
[0012] The present invention is further described below in conjunction with specific embodiments.
[0013] Embodiment A method for extracting water bodies by segmentation based on remote sensing of a watershed based on terrain information comprises the following steps: S1. Collect DEM (digital elevation model) data, building land vector data, and basin optical remote sensing image data information of the target basin: Take a river basin in Shandong Province as the research area, collect the digital elevation model (DEM) data, building land vector data, Sentinel-2 optical remote sensing image data and other information of the river basin, as follows: Figure 2 , 3 , as shown in Figure 4.
[0014] S2. Use ArcGIS software to calculate the slope of the collected DEM data to obtain the slope value S : The collected DEM data was used to calculate the slope using the surface analysis tool under the Spatial Analyst tool in ArcGIS software. The calculation results are as follows: Figure 5 shown.
[0015] S3. Calculate the normalized digital elevation model value, using the normalized digital elevation model value and slope value S Constructing the watershed division index WSDI based on terrain changes: According to the digital elevation model value within the watershed, the normalized digital elevation model value is calculated. The calculation formula is as follows: , In the formula, ND_DEM is the normalized digital elevation model value; DEM is the digital elevation model value, in meters; DEM max It is the maximum value of the digital elevation model within the watershed, in meters; DEM minIt is the minimum value of the digital elevation model within the watershed, in meters; The watershed division index (WDI) based on terrain changes was constructed by integrating the normalized digital elevation model values and slope. WSDI ), the calculation formula is as follows: , In the formula, WSDI is the watershed division index, ND_DEM is the normalized digital elevation model value, S is the slope value; From the DEM data, we can know that the maximum value of a river basin DEM ( DEM max ) is 930 m, and the minimum value of DEM (; DEM min ) is -15 meters, calculate ND_DEM , the results are as follows Figure 6 As shown; The normalized digital elevation model values and slopes are integrated to construct a watershed division index based on terrain changes. The results are as follows: Figure 7 shown.
[0016] S4. Divide the basin into three areas: upper, middle and lower reaches according to the value range of the basin division index: The classification criteria are as follows: , Where: a , b is a constant, which is related to the watershed where it is located; The division standard of a river basin is as follows: , The results of watershed division are as follows Figure 8 shown.
[0017] S5. Based on the construction land vector data, the middle and downstream areas are further divided into construction areas and non-construction areas. Combined with the watershed division and construction land vector data, the water body extraction area is divided into upstream and midstream construction areas, midstream non-construction areas, downstream construction areas, and downstream non-construction areas: According to the vector data of building land, the middle and downstream areas are further divided into building area and non-building area. The division is shown in the following formula. Different water body extraction methods are used in the building area and non-building area. The division results are shown in Figure 9 , .
[0018] S6. Select different water extraction algorithms for different water extraction areas, and use the shadow water index in the upper reaches of the basin. SWI , the modified normalized difference water index was selected in the built-up areas in the middle and lower reaches of the basin MNDVI , select the normalized water index in the non-built-up areas in the middle and lower reaches of the basin NDVI : The shadows of mountains and bare land in the upper reaches of the basin are the main factors that interfere with the extraction of water bodies, so the shadow water index ( SWI ), in the middle and lower reaches of the basin, the shadow of buildings is the main factor interfering with the extraction of water bodies, so the water bodies in the urban area are selected by the modified normalized difference water index ( MNDVI ), the normalized water index ( NDVI ), as shown below: , , , In the formula, B is the reflectance value of the blue band, G is the reflectance value of the green band, NIR is the reflectance value in the near infrared band, WORLD is the reflectivity value in the mid-infrared band.
[0019] S7. Comprehensively construct a comprehensive regional remote sensing water extraction method based on watershed division, building area division, and water index selection, and determine the segmented water extraction algorithm for the entire watershed: , In the formula, B is the reflectance value of the blue band, G is the reflectance value of the green band, NIR is the reflectance value in the near infrared band, WORLDis the reflectance value of the mid-infrared band, n1, n2, n3, n4, and n5 are the water extraction thresholds, and the threshold values are determined according to the experience of the region and the actual situation. Taking a river basin as an example, the value selection process of n1, n2, n3, n4, and n5 is shown respectively. Among them: the experience of n1, n2, n3, n4, and n5 is generally 0.08. 0.08 is used as the threshold to extract water bodies from images of different regions using the corresponding extraction algorithm. The confusion matrix method is used to evaluate the accuracy, and the classification results are verified for accuracy. If the overall accuracy is greater than 85%, the threshold is used; if the overall accuracy is less than 85%, and the omission error is greater than the misclassification error, it means that the water body is missed, and it is decreased by 0.01 as a step size, and the accuracy evaluation is continued after the decrease until it meets the accuracy standard; if the overall accuracy is less than 85%, and the misclassification error is greater than the omission error, it means that the water body is wrongly extracted, and it is increased by 0.01 as a step size, and the accuracy evaluation is continued after the increase until it meets the accuracy standard.
[0020] Combined with the actual situation of the study area, the water extraction formula is formed as follows: , Where: B is the reflectance value of the blue band, G is the reflectance value of the green band, NIR is the reflectance value in the near infrared band, WORLD is the reflectivity value in the mid-infrared band.
[0021] S8. For the water bodies in the basin, the formula obtained in step S7 is used to extract the water bodies in different regions. The results of the water bodies in a certain river basin are as follows: Figure 10 shown.
[0022] The above is a detailed introduction of the present invention in combination with specific embodiments, and the protection scope of the present invention is not limited thereto.
Claims
1. A method for extracting water bodies by segmentation based on remote sensing of a watershed based on terrain information, characterized in that: The steps include: S1. Collect DEM data, construction land vector data, and optical remote sensing image data of the target watershed; S2. Use ArcGIS software to calculate the slope of the collected DEM data to obtain the slope value S ; S3. Calculate the normalized digital elevation model value, using the normalized digital elevation model value and slope value S Constructing a watershed division index based on topographic changes WSDI ; S4. Divide the basin into three areas: upper, middle and lower reaches according to the value range of the basin division index; S5. The middle and lower reaches are further divided into built-up areas and non-built-up areas according to the built-up land vector data. The water body extraction area is divided into the upper reaches, the middle reaches built-up areas, the middle reaches non-built-up areas, the lower reaches built-up areas, and the lower reaches non-built-up areas in combination with the watershed division and the built-up land vector data; S6. Select different water extraction algorithms for different water extraction areas, and use the shadow water index in the upper reaches of the basin. SWI , the modified normalized difference water index was selected in the built-up areas in the middle and lower reaches of the basin MNDWI , select the normalized water index in the non-built-up areas in the middle and lower reaches of the basin NDWI ; S7. Comprehensively construct a comprehensive regional remote sensing water extraction method based on watershed division, building area division, and water index selection, and determine the segmented water extraction algorithm for the entire watershed: ， In the formula, WSDI is the watershed division index, a , b The threshold for watershed division, C is a built-up area or a non-built-up area, n1, n2, n3, n4, and n5 are water extraction thresholds, and the threshold values are determined based on the experience of the area and the actual situation; S8. For the water bodies within the watershed, perform regional extraction according to the formula obtained in step S7.
2. The method for extracting water bodies by segmentation based on remote sensing of a river basin based on terrain information according to claim 1, characterized in that: The calculation formulas for the normalized digital elevation model value and watershed division index described in step S3 are as follows: ， In the formula, ND_DEM is the normalized digital elevation model value, DEM is the digital elevation model value, DEM max is the maximum value of the digital elevation model within the watershed. DEM min It is the minimum value of the digital elevation model within the watershed; ， In the formula, WSDI is the watershed division index, ND_DEM is the normalized digital elevation model value, S is the slope value.
3. The method for extracting water bodies by segmentation based on topographic information from remote sensing of a watershed according to claim 1, characterized in that: The watershed division criteria described in step S4 are as follows: ， In the formula, a , b is a constant and is related to the watershed in which it is located.
4. The method for extracting water bodies by segmentation based on remote sensing of a river basin based on terrain information according to claim 1, characterized in that: In step S5, the construction area and non-construction area are divided as follows: 。 5. The method for extracting water bodies by segmentation based on remote sensing of a river basin based on terrain information according to claim 1, characterized in that: The calculation formulas for each water body index in step S6 are as follows: ， ， ， In the formula, B is the reflectance value of the blue band, G is the reflectance value of the green band, NIR is the reflectance value in the near infrared band, MIR is the reflectivity value in the mid-infrared band.
6. The method for extracting water bodies by segmentation based on remote sensing of a river basin based on terrain information according to claim 5, characterized in that: The final water body extraction algorithm in step S7 is: ， In the formula, B is the reflectance value of the blue band, G is the reflectance value of the green band, NIR is the reflectance value in the near infrared band, MIR is the reflectance value in the mid-infrared band, n1, n2, n3, n4, and n5 are the water extraction thresholds, and the threshold values are determined based on the empirical values of the area and the actual situation.
7. The method for extracting water bodies by segmentation based on remote sensing of a river basin based on terrain information according to claim 1, characterized in that: For the values of n1, n2, n3, n4, and n5 in step S7, first take the empirical value as the threshold to extract water bodies from images of different areas using the corresponding extraction algorithm, use the confusion matrix method to evaluate the accuracy, and verify the accuracy of the classification results. If the overall accuracy is greater than 85%, the threshold is used; if the overall accuracy is less than 85%, and the omission error is greater than the misclassification error, it means that the water body is missed, and the accuracy is decreased by 0.01 in steps, and the accuracy is continued after the decrease until it meets the accuracy standard; if the overall accuracy is less than 85%, and the misclassification error is greater than the omission error, it means that the water body is wrongly extracted, and the accuracy is increased by 0.01 in steps, and the accuracy is continued after the increase until it meets the accuracy standard.