Reservoir hydro-fluctuation belt range identification and extraction method of satellite remote sensing image

Through satellite remote sensing images and automatic water extraction index AWEIsh, combined with Otsu algorithm, the problems of low efficiency and poor accuracy of traditional reservoir desolation belt extraction methods are solved, and efficient and accurate identification of desolation belt range is achieved, providing reservoir ecological environment Scientific support.

CN118941943APending Publication Date: 2024-11-12BEIJING FORESTRY UNIVERSITY
View PDF 0 Cites 1 Cited by

Patent Information

Application Number
CN202410928292.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-11
Publication Date
2024-11-12

AI Technical Summary

Technical Problem

The traditional reservoir elimination belt extraction method relies on ground investigation and manual interpretation, which is very labor-intensive, inefficient and difficult to ensure accuracy.

Method used

Satellite remote sensing images are used to obtain high-score images, and the optimal threshold is determined through the automatic water extraction index AWEIsh and Otsu algorithms, the water body area is automatically identified, and the water body area changes are calculated to determine the range of the desolation zone.

Benefits of technology

It improves the identification efficiency and accuracy of the reservoir desolation zone range and provides scientific data to support reservoir ecological environment assessment and governance.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118941943A_ABST
    Figure CN118941943A_ABST
Patent Text Reader

Abstract

The invention discloses a reservoir hydro-fluctuation belt range identification and extraction method for a satellite remote sensing image, and the method comprises the steps: obtaining a high-resolution remote sensing image of each month in a required time period of a target region, and carrying out the preprocessing of the high-resolution remote sensing image, and obtaining the reflectivity values of different spectral bands of the remote sensing image; according to the obtained reflectivity values of different spectral bands, calculating an automatic water body extraction index of a required time period of the target area; determining an optimal threshold value of automatic water body extraction index segmentation; segmenting a water body region according to the automatic water body extraction index and the optimal threshold value; comparing according to the area of the water body region to determine the minimum and maximum time periods of the area of the water body region, wherein the water body region in the time period corresponds to the low water level and the flood level; overlapping the low-level water body area and the flood-level water body area, and determining a low-level water body change area and a flood-level water body change area, namely a reservoir hydro-fluctuation belt; the technical problems that in the prior art, a reservoir hydro-fluctuation belt extraction method mainly depends on ground survey and manual interpretation, the workload is large, and the efficiency is low are solved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention belongs to the technical field of ecological environment monitoring, and in particular relates to a method for identifying and extracting the range of a reservoir drawdown zone from satellite remote sensing images. Background Art

[0002] The drawdown zone of a reservoir refers to the area covered alternately by land and water due to the change of the reservoir water level between low water level and flood level. The ecological environment of the drawdown zone is complex and changeable, and is affected by water level changes all year round, which can easily cause problems such as soil erosion and water pollution. Therefore, accurately identifying and extracting the scope of the drawdown zone of a reservoir is of great scientific significance for ecological restoration and governance.

[0003] The traditional method of extracting reservoir drawdown zone mainly relies on ground survey and manual interpretation, which is labor-intensive, inefficient and difficult to ensure accuracy. Summary of the invention

[0004] The technical problem to be solved by the present invention is to provide a method for identifying and extracting the range of a reservoir drawdown zone from satellite remote sensing images, so as to solve the technical problems that the existing reservoir drawdown zone extraction method mainly relies on ground investigation and manual interpretation, has a large workload, low efficiency, and is difficult to ensure accuracy.

[0005] The technical solution of the present invention is:

[0006] A method for identifying and extracting the range of a reservoir drawdown zone from satellite remote sensing images, the method comprising:

[0007] Step 1: Obtain high-resolution remote sensing images of the target area every month within the required time period, and obtain the reflectance values ​​of different spectral bands of the remote sensing images after preprocessing;

[0008] Step 2: Calculate the automatic water extraction index AWEI for the required time period of the target area based on the reflectance values ​​of different spectral bands obtained sh ;

[0009] Step 3: Determine the Automatic Water Extraction Index AWEI sh Optimal threshold for segmentation;

[0010] Step 4: According to the automatic water extraction index AWEI sh and optimal threshold segmentation of water areas, AWEI sh Areas with >t are identified as water areas;

[0011] Step 5: Compare the water area extracted each month within the required time period to determine the time period with the smallest and largest water area. The water area in this period is the water area corresponding to the low water level and flood level; superimpose the low water level and flood level water area to determine the low water level and flood level water change area, which is the reservoir drawdown zone.

[0012] Preprocessing includes radiation correction, atmospheric correction and geometric correction of the acquired remote sensing image data.

[0013] Automatic Water Extraction Index AWEI sh The calculation formula is as follows:

[0014] AWEI sh =1.2×ρ band1 +2.2×ρ band2 -1.5×ρ band4 -2.0×ρ band5 -0.15

[0015] ×ρ band7

[0016] Where: ρ is the reflectance value of different spectral bands; band1 is the wavelength captured by the blue light band; band2 is the wavelength captured by the green light band; band4 is the wavelength captured by the near-infrared band; band5 is the wavelength captured by the short-wave infrared band; band7 is the wavelength captured by the short-wave infrared band.

[0017] The optimal threshold is determined using the Otsu algorithm, which includes:

[0018] Step 3.1. Calculate AWEI sh The histogram of the values ​​is obtained for each AWEI sh The frequency with which the value occurs is then normalized to a probability distribution so that the sum is 1;

[0019] Let L be the number of gray levels (usually 256), and the probability distribution calculation formula of P(i) is:

[0020]

[0021] Where n i is the number of pixels at gray level i; N is the total number of pixels in the image;

[0022] Step 3.2: Calculate from AWEI sh Cumulative probability ω(t) and cumulative mean μ(t) from the minimum value to the threshold t:

[0023]

[0024] Calculate the mean μ of the entire image T

[0025]

[0026] Step 3.3, calculate the between-class variance, between-class variance The formula is as follows

[0027]

[0028] Wherein, ω(t) is the proportion of pixels to the left of threshold t (including t) to the total pixels; μ(t) is the average grayscale value of pixels to the left of threshold t (including t); μ T is the average gray value of the entire image;

[0029] Step 3.4: Screen all possible thresholds t and find the one that makes the inter-class variance The largest threshold t is the optimal threshold.

[0030] The output water body of the water area is in binary array raster data format, where pixels in the water area are marked as 1 and pixels in the non-water area are marked as 0.

[0031] The method for extracting the drawdown zone of a reservoir includes: using spatial analysis software to add water body raster layers of low water level and flood level to the map, opening the raster calculation tool, and using the water body raster layer of the flood level minus the water body raster layer of the low water level in the raster calculator to generate a new raster layer, in which pixels with a value of 1 represent the water body area newly added at the flood level compared to the low water level, that is, the drawdown zone, and pixels with a value of 0 represent areas that have not changed.

[0032] Beneficial effects of the present invention:

[0033] The present invention uses high-resolution remote sensing images to extract water areas and determines the range of the drawdown zone based on the water body change area, which is highly efficient. Satellite remote sensing images can be used to quickly obtain water body change information in a large reservoir area, which significantly improves work efficiency compared to traditional ground survey methods. By calculating the automatic water body extraction index and the optimal threshold, it can accurately distinguish water bodies from other land types, ensuring the accuracy of the recognition of the drawdown zone range. It provides scientific drawdown zone range data and provides effective support for reservoir ecological environment assessment and management.

[0034] The invention solves the technical problems that the existing method for extracting the drawdown zone of a reservoir mainly relies on ground investigation and manual interpretation, has a large workload, low efficiency, and is difficult to ensure accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 It is a schematic diagram of the process of the present invention;

[0036] Figure 2 Automatic Water Extraction Index AWEI for specific implementation example sh Schematic diagram of calculation results;

[0037] Figure 3 This is a schematic diagram of water distribution in a specific implementation example;

[0038] Figure 4 It is a schematic diagram of the distribution of the drawdown zones of the implementation case in the specific implementation method;

[0039] Figure 5 It is a schematic diagram of the calculation results of NDWI and MNDWI of the implementation case in the specific implementation manner;

[0040] Figure 6 This is a schematic diagram of the water area distribution extracted under the NDWI and MNDWI empirical values ​​of the implementation case in the specific implementation method. DETAILED DESCRIPTION

[0041] like Figure 1 The figure shows a method for identifying and extracting the range of the reservoir drawdown zone from satellite remote sensing images, which includes the following steps:

[0042] S1: Obtain the longitude and latitude information of the boundary control points of the target area, which can be located using various map software; obtain high-resolution remote sensing images of the target area every month within the required time period, and specifically obtain Sentinel open source remote sensing images through the Google Earth Engine website. Perform preprocessing such as radiation correction, atmospheric correction, and geometric correction on the acquired remote sensing image data, which can be performed using software such as ENVI or ERDAS to obtain the reflectance values ​​of different spectral bands of the remote sensing image.

[0043] S2: Based on the reflectance values ​​of different spectral bands, use ENVI or ERDAS and other software to calculate the automatic water extraction index AWEI for the required time period of the target area through its band operation function sh , the calculation formula is as follows:

[0044] AWEI sh =1.2×ρ band1 +2.2×ρ band2 -1.5×ρ band4 -2.0×ρ band5 -0.15

[0045] ×ρ band7

[0046] Where ρ is the reflectance value of different spectral bands. Band1 (blue light band): the captured wavelength range is about 0.45-0.52 microns; band2 (green light band): the captured wavelength range is about 0.52-0.60 microns; band4 (near infrared band): the captured wavelength range is about 0.76-0.90 microns. Band5 (short-wave infrared band): the captured wavelength range is about 1.55-1.75 microns. Band7 (short-wave infrared band): the captured wavelength range is about 2.08-2.35 microns.

[0047] In the past, water extraction methods mostly used the normalized difference water index NDWI and the improved normalized difference water index MNDWI to extract water. The normalized difference water index NDWI is easily affected by vegetation and buildings, and the error is large in areas with high vegetation coverage or in urban areas. The improved normalized difference water index MNDWI may be misclassified in complex environments. Automatic Water Extraction Index AWEI sh By using the reflectance information of multiple bands and weighted combination, the ability to distinguish between water bodies and non-water bodies is improved. At the same time, the reflectance characteristics of different bands are taken into account, which can maintain high robustness under various environmental conditions, especially under complex geographical and climatic conditions. The automatic water extraction index AWEI sh It can effectively distinguish water bodies from other landforms. Compared with the previous methods of normalized difference water index NDWI and improved normalized difference water index MNDWI, automatic water extraction index AWEI sh It performs well in complex environments, with better applicability and higher extraction quality.

[0048] S3: Determine the automatic water extraction index AWEI using the Otsu algorithm sh The best threshold for segmentation. In the past, water extraction often used the empirical value of 0 as the segmentation threshold, while the Otsu algorithm is an algorithm that can automatically determine the image binarization threshold, aiming to maximize the inter-class variance to best separate the foreground and background. This method not only reduces the subjectivity of manually setting the threshold, but also improves the accuracy and robustness of water extraction. The process of calculating the optimal threshold t by the Otsu algorithm is as follows:

[0049] (1) Calculate histogram and probability distribution

[0050] First, calculate AWEI sh The histogram of the values ​​is obtained for each AWEI sh The frequencies of the values ​​are then normalized to a probability distribution so that they sum to 1.

[0051] Let L be the number of gray levels (usually 256), and the probability distribution calculation formula of P(i) is:

[0052]

[0053] Where n i is the number of pixels at gray level i; N is the total number of pixels in the image.

[0054] (2) Calculate the cumulative probability and mean

[0055] Calculation from AWEI sh Cumulative probability ω(t) and cumulative mean μ(t) from the minimum value to the threshold t:

[0056]

[0057] Calculate the mean μ of the entire image T

[0058]

[0059] (3) Calculate the inter-class variance

[0060] Between-class variance The formula is as follows

[0061]

[0062] Wherein, ω(t) is the proportion of pixels to the left of threshold t (including t) to the total pixels; μ(t) is the average grayscale value of pixels to the left of threshold t (including t); μ T is the average gray value of the entire image.

[0063] (4) Finding the optimal threshold

[0064] Screen all possible thresholds t and find the one that makes the between-class variance The largest threshold t is the optimal threshold.

[0065] S4: Automatic water extraction index AWEI obtained using S2 and S3 sh and the optimal threshold t to segment the water area, AWEI sh The area with a value greater than t is identified as a water body area, and the output water body is a binary array (raster data format), where the pixels in the water body area are marked as 1 and the pixels in the non-water body area are marked as 0. The automatic water extraction index AWEI sh Segmenting water bodies with the optimal threshold t can significantly improve the accuracy and efficiency of water body extraction. It has the characteristics of automation, robustness and strong applicability, making it have broad application prospects in water resources management, environmental monitoring and disaster warning.

[0066] S5: Compare the water area extracted each month within the required period to determine the period with the smallest and largest water area. The water area in this period is the water area corresponding to the low water level and flood level. Overlay the low water level and flood level water area to determine the area where the water body changes at the low water level and flood level, which is the reservoir drawdown zone. Specifically, ArcGIS spatial analysis software can be used to add the water body raster layers of the low water level and flood level to the map, open the raster calculator tool, and use the water body raster layer of the flood level to subtract the water body raster layer of the low water level in the raster calculator to generate a new raster layer, in which pixels with a value of 1 represent the water body area that is newly added to the flood level compared to the low water level, that is, the drawdown zone. Pixels with a value of 0 are areas that have not changed.

[0067] The technical solution of the present invention is described below by taking examples:

[0068] Taking a reservoir in the alpine canyon area of ​​southwest China from April to September as an example, the technical solution of the present invention is explained:

[0069] S1: Obtain the boundary of the target area, use the Google Earth Engine website to obtain the high-resolution remote sensing images of the low water level and flood level of the target area, and perform preprocessing such as radiation correction, atmospheric correction and geometric correction through ENVI software to obtain the reflectance data of the remote sensing image.

[0070] S2: Based on the reflectance data of the remote sensing image obtained in S1, the automatic water extraction index AWEI is calculated by using ENVI software through its band operation function sh , the calculation formula is as follows: AWEI sh =1.2×ρ band1 +2.2×ρ band2 -1.5×ρ band4 -2.0×ρ band5 -0.15

[0071] ×ρ band7

[0072] Where ρ is the reflectance value of different spectral bands. Band1 (blue light band): the captured wavelength range is about 0.45-0.52 microns; band2 (green light band): the captured wavelength range is about 0.52-0.60 microns; band4 (near infrared band): the captured wavelength range is about 0.76-0.90 microns. Band5 (short-wave infrared band): the captured wavelength range is about 1.55-1.75 microns. Band7 (short-wave infrared band): the captured wavelength range is about 2.08-2.35 microns.

[0073] Figure 2 Automatic Water Extraction Index AWEI sh Schematic diagram of calculation results.

[0074] S3: Automatic Water Extraction Index AWEI calculated in S2 sh , the Otsu algorithm is used to determine the optimal threshold for normalized water index segmentation. The calculation process is as follows:

[0075] (1) Calculate histogram and probability distribution

[0076] First, calculate AWEI sh The histogram of the values ​​is obtained for each AWEI sh The frequencies of the values ​​are then normalized to a probability distribution so that they sum to 1.

[0077] Let L be the number of gray levels (usually 256), and the probability distribution calculation formula of P(i) is:

[0078]

[0079] Where n i is the number of pixels at gray level i; N is the total number of pixels in the image.

[0080] (2) Calculate the cumulative probability and mean

[0081] Calculate the cumulative probability ω(t) and cumulative mean μ(t) from the minimum value of AWEIsh to the threshold t:

[0082]

[0083] Calculate the mean μ of the entire image T

[0084]

[0085] (3) Calculate the inter-class variance

[0086] Between-class variance The formula is as follows

[0087]

[0088] Wherein, ω(t) is the proportion of pixels to the left of threshold t (including t) to the total pixels; μ(t) is the average grayscale value of pixels to the left of threshold t (including t); μ T is the average gray value of the entire image.

[0089] (4) Finding the optimal threshold

[0090] Screen all possible thresholds t and find the one that makes the between-class variance The largest threshold t is the optimal threshold. In this example, the optimal thresholds extracted in April, May, June, July, August and September are 0.017, 0.015, 0.015, 0.020, 0.020 and 0.016 respectively.

[0091] S4: Automatic water extraction index AWEI obtained using S2 and S3 sh and the optimal threshold t to segment the water area, AWEI sh The area > t is identified as the water area, and the output water body is a binary array (raster data format), where the pixels in the water area are marked as 1 and the pixels in the non-water area are marked as 0.

[0092] Figure 3 This is a schematic diagram of the water body regional distribution.

[0093] S5: Compare the water area extracted each month within the required time period to determine the time period with the smallest and largest water area. The water area in this period is the water area corresponding to the low water level and flood level (the corresponding time periods in the example are May and August, respectively). Overlay the water area of ​​the low water level and flood level to determine the area where the water body changes at the low water level and flood level, which is the reservoir drawdown zone. Specifically, ArcGIS spatial analysis software can be used to add the water body raster layers of the low water level and flood level to the map, open the raster calculation tool, and use the water body raster layer of the flood level minus the water body raster layer of the low water level in the raster calculator to generate a new raster layer, in which pixels with a value of 1 represent the water body area that is newly added to the flood level compared to the low water level, that is, the drawdown zone. Pixels with a value of 0 are areas that have not changed.

[0094] Figure 4 This is a schematic diagram of the distribution of the ebb and flow zone.

[0095] Water extraction is the basis of ebb and flow zone extraction. We compared the automatic water extraction index (AWEI) sh The results of water extraction in August in the required period using the Normalized Difference Water Index (NDWI) and the Improved Normalized Difference Water Index (MNDWI). The Normalized Difference Water Index (NDWI) and the Improved Normalized Difference Water Index (MNDWI) are compared with the Automatic Water Extraction Index (AWEI). sh It is more susceptible to the shadows of buildings and mountains, and large areas of water bodies are incorrectly extracted during water extraction, while the automatic water extraction index AWEI sh The obtained water body areas have a high classification consistency with the real water body areas.

[0096] Figure 5 Schematic diagram of NDWI and MNDWI calculation results.

[0097] Figure 6 Schematic diagram of water area distribution extracted under the empirical values ​​of NDWI and MNDWI.

Claims

1. A method for identifying and extracting the range of a reservoir drawdown zone from satellite remote sensing images, characterized in that: The method comprises: Step 1: Obtain high-resolution remote sensing images of the target area every month within the required time period, and obtain the reflectance values ​​of different spectral bands of the remote sensing images after preprocessing; Step 2: Calculate the automatic water extraction index AWEI for the required time period of the target area based on the reflectance values ​​of different spectral bands obtained sh ; Step 3: Determine the Automatic Water Extraction Index AWEI sh Optimal threshold for segmentation; Step 4: According to the automatic water extraction index AWEI sh and optimal threshold segmentation of water areas, AWEI sh Areas with >t are identified as water areas; Step 5: Compare the water area extracted each month within the required time period to determine the time period with the smallest and largest water area. The water area in this period is the water area corresponding to the low water level and flood level; superimpose the low water level and flood level water area to determine the low water level and flood level water change area, which is the reservoir drawdown zone.

2. The method for identifying and extracting the range of the reservoir drawdown zone from satellite remote sensing images according to claim 1 is characterized by: Preprocessing includes radiation correction, atmospheric correction and geometric correction of the acquired remote sensing image data.

3. The method for identifying and extracting the range of the reservoir drawdown zone from satellite remote sensing images according to claim 1 is characterized by: Automatic Water Extraction Index AWEI sh The calculation formula is as follows: AWEI sh =1.2×ρ band1 +2.2×ρ band2 -1.5×ρ band4 -2.0×ρ band5 -0.15×ρ band7 Where: ρ is the reflectance value of different spectral bands; band1 is the wavelength captured by the blue light band; band2 is the wavelength captured by the green light band; band4 is the wavelength captured by the near-infrared band; band5 is the wavelength captured by the short-wave infrared band; band7 is the wavelength captured by the short-wave infrared band.

4. The method for identifying and extracting the range of the reservoir drawdown zone from satellite remote sensing images according to claim 1 is characterized by: The optimal threshold is determined using the Otsu algorithm, which includes: Step 3.

1. Calculate AWEI sh The histogram of the values ​​is obtained for each AWEI sh The frequency with which the value occurs is then normalized to a probability distribution so that the sum is 1; Let L be the number of gray levels (usually 256), and the probability distribution calculation formula of P(i) is: where n i is the number of pixels at gray level i; N is the total number of pixels in the image; Step 3.2: Calculate from AWEI sh Cumulative probability ω(t) and cumulative mean μ(t) from the minimum value to the threshold t: Calculate the mean μ of the entire image T Step 3.3, calculate the between-class variance, between-class variance The formula is as follows Wherein, ω(t) is the proportion of pixels to the left of threshold t (including t) to the total pixels; μ(t) is the average grayscale value of pixels to the left of threshold t (including t); μ T is the average gray value of the entire image; Step 3.4: Screen all possible thresholds t and find the one that makes the inter-class variance The largest threshold t is the optimal threshold.

5. The method for identifying and extracting the range of the reservoir drawdown zone from satellite remote sensing images according to claim 1 is characterized by: The output water body of the water area is in binary array raster data format, where pixels in the water area are marked as 1 and pixels in the non-water area are marked as 0.

6. The method for identifying and extracting the range of the reservoir drawdown zone from satellite remote sensing images according to claim 1, characterized in that: The method for extracting the drawdown zone of a reservoir includes: using spatial analysis software to add water body raster layers of low water level and flood level to the map, opening the raster calculation tool, and using the water body raster layer of the flood level minus the water body raster layer of the low water level in the raster calculator to generate a new raster layer, in which pixels with a value of 1 represent the water body area newly added at the flood level compared to the low water level, that is, the drawdown zone, and pixels with a value of 0 represent areas that have not changed.

Citation Information

Cited By

  • Multispectral remote sensing image shallow water region extraction method based on object adaptive NDWI threshold

    CN120707581A