A water area extraction method fusing radar backscattering and interferometric coherence

Abstract

The application discloses a water area extraction method fusing radar backscattering and interference coherence coefficients, and belongs to the technical field of remote sensing water area extraction, and comprises the following steps: S1, acquiring a polarimetric SAR image, and pre-processing the polarimetric SAR image to obtain a backscattering coefficient image and an interference coherence coefficient image; S2, calculating an SDWI water body index according to the backscattering coefficient image; S3, determining a water body preliminary extraction result according to the SDWI water body index; S4, performing fusion processing on the water body preliminary extraction result based on the interference coherence coefficient image to generate a fused water body extraction result; and S5, performing refinement processing on the fused water body extraction result to obtain a final water body range. The application can not only accurately calculate a segmentation threshold, but also can avoid the influence of submerged plants and silt on the extraction result, and improves the accuracy of water area extraction.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of remote sensing water area extraction, and particularly relates to a water area extraction method fusing radar backscattering and interference coherence coefficients. BACKGROUND

[0002] Lakes, as an important part of water resources, play a crucial role in the water cycle in nature. Lakes not only participate in processes such as climate regulation and water conservation, but also are essential to maintaining the balance of the ecosystem. Lake area is one of the important indicators to measure water level, and accurate extraction and monitoring of lake area has important scientific significance and application value. With the development of earth observation technology, remote sensing is widely used for water area extraction and has gradually become an important technical means for lake area extraction. Compared with optical remote sensing, Sentinel-1 SAR satellite has the advantages of all-weather, all-day, cloud and rain penetration, and rapid data acquisition, providing an opportunity for high-precision water body identification and area extraction. Among them, threshold segmentation method is one of the main methods for extracting water area using remote sensing data, mainly including manual visual selection of threshold and automatic threshold selection. The threshold selected by manual visual selection has strong subjectivity, and cannot accurately divide water body and non-water body, and the method is time-consuming and difficult to meet the application under the condition of large-scale and massive data. For the division of water body and non-water body, the commonly used automatic threshold selection is OTSU algorithm, which has the characteristics of simple operation and quick use. However, when the automatic threshold selection method is used for water body and non-water body identification based on SAR image, it is easily affected by factors such as mountain shadow and marsh caused by SAR side-looking imaging, resulting in misclassification. In addition, the current water body extraction method based on Sentinel-1 SAR data mostly uses single polarization information, and due to the influence of submerged plants and silt, the water body reflectivity and the reflectivity of submerged plants and silt show high similarity in the backscattering coefficient image, resulting in reduced water body area extraction accuracy. SUMMARY

[0003] The application is proposed to solve the above problems, and provides a water area extraction method fusing radar backscattering and interference coherence coefficients. The beneficial effects of the further scheme are as follows:

[0004] The technical scheme of the application is a water area extraction method fusing radar backscattering and interference coherence coefficients, which comprises the following steps:

[0005] S1, acquiring a polarized SAR image and pre-processing the polarized SAR image to obtain a backscattering coefficient image and an interference coherence coefficient image;

[0006] S2, calculating an SDWI water body index according to the backscattering coefficient images of different polarization modes;

[0007] S3, determining a preliminary water extraction result according to the SDWI water body index by using an iterative threshold segmentation method;

[0008] S4, performing fusion processing on the preliminary water extraction result based on the interference coherence coefficient image to generate a fused water extraction result;

[0009] S5, performing refinement processing on the fused water extraction result to obtain a final water body range.

[0010] Further, in S1, the polarimetric SAR image is preprocessed by orbit correction, radiation correction, TOPS SAR image declination, cropping, multi-view, DN value conversion to dB, and geographic coding to obtain a backscattering coefficient image.

[0011] In S1, the polarimetric SAR image is preprocessed by orbit correction, radiation correction, TOPS SAR image segmentation, reverse geographic coding, TOPS SAR image declination, TOPS SAR image splicing, multi-view, interference, cropping, and geographic coding to obtain an interference coherence coefficient image.

[0012] Further, in S2, the SDWI water body index K SDWI is calculated according to the following formula:

[0013] K SDWI = ln(10*VV*VH);

[0014] In the formula, VV represents the backscattering coefficient in the vertical polarization mode, and VH represents the backscattering coefficient in the cross-polarization mode.

[0015] Further, S3 includes the following sub-steps:

[0016] S31, dividing the gray-scale image corresponding to the SAR image into a plurality of gray-scale levels and calculating the probability of each gray-scale level;

[0017] S32, calculating the background class probability and the foreground class probability according to the probability of each gray-scale level;

[0018] S33, calculating the average gray-scale value of the background class and the average gray-scale value of the foreground class according to the background class probability and the foreground class probability;

[0019] S34, calculating the variance between different classes according to the background class probability, the foreground class probability, the average gray-scale value of the background class, and the average gray-scale value of the foreground class;

[0020] S35, calculating the variance between different elements in the background class and the variance between different elements in the foreground class according to the background class probability, the foreground class probability, the average gray-scale value of the background class, and the average gray-scale value of the foreground class;

[0021] S36, calculating the ratio of the sum of inter-class variance and intra-class variance according to the variance between different classes, the variance between different elements in the background class, and the variance between different elements in the foreground class;

[0022] S37, repeating S32-S36 to obtain the ratio of the sum of inter-class variance and intra-class variance of each gray value, and extracting the gray value corresponding to the maximum ratio;

[0023] S38, calculating the sum of entropy of each gray value, and extracting the gray value corresponding to the maximum sum of entropy;

[0024] S39, calculating the division threshold according to the gray value corresponding to the maximum ratio and the gray value corresponding to the maximum sum of entropy;

[0025] S310, determining the preliminary extraction result of water body according to the division threshold and the SDWI water body index.

[0026] Further, in S31, the probability P i of the i-th gray level is calculated according to the following formula:

[0027]

[0028] In the formula, n i represents the number of pixels with gray value i in the image, and N represents the total number of pixels of the gray image;

[0029] In S32, the calculation formula of the background class probability P a (k) is:

[0030]

[0031] In the formula, k represents the division threshold of the gray image;

[0032] In S32, the calculation formula of the foreground class probability P b (k) is:

[0033]

[0034] In the formula, L represents the number of gray levels of the gray image;

[0035] In S33, the calculation formula of the average gray value m a (k) of the background class is:

[0036]

[0037] In S33, the calculation formula of the average gray value m b (k) of the foreground class is:

[0038]

[0039] In S34, the variance σ0(k) between different classes 2 The calculation formula of σ0(k) is:

[0040] σ0(k) 2 = P a (k)P b (k)(m a (k)-m b (k)) 2 ;

[0041] In S35, the variance σ a The calculation formula of σ

[0042]

[0043] In S35, the variance σ b The calculation formula of σ

[0044]

[0045] In S36, the calculation formula of the ratio θ of the sum of the inter-class variance and the intra-class variance is:

[0046]

[0047] In S38, the calculation formula of the sum λ of the entropies of the gray values is:

[0048]

[0049] In the formula, ln(·) represents a logarithmic function;

[0050] In S39, the calculation formula of the division threshold T1 is:

[0051]

[0052] In the formula, t a represents the gray value corresponding to the maximum ratio, and t b represents the gray value corresponding to the maximum sum of entropies.

[0053] Further, S310 includes the following sub-steps:

[0054] S3101, dividing into a background region and a foreground region by using the division threshold;

[0055] S3102, respectively weighting and averaging the background and foreground regions based on the SDWI water body index to obtain a weighted average gray value of the background region and a weighted average gray value of the foreground region;

[0056] S3103, pixels greater than the weighted average gray value of the background region and less than the weighted average gray value of the foreground region are taken as uncertain regions;

[0057] S3104, the division threshold of the uncertain region is iteratively calculated until the difference between the final division threshold and the division threshold of the last iteration is less than the convergence threshold, and the final division threshold is taken as the best threshold;

[0058] S3105, the region with a gray value lower than the best threshold in the backscattering coefficient image is taken as the preliminary water body result.

[0059] Further, in S3102, the weighted average gray value u a of the background region is calculated according to the following formula:

[0060]

[0061] In the formula, T1 represents the division threshold, w(i) represents the weight of the pixel, f(x,y) represents the SDWI water body index image, x represents the horizontal coordinate of each pixel on the two-dimensional remote sensing image, and y represents the vertical coordinate of each pixel on the two-dimensional remote sensing image.

[0062] In S3102, the weighted average gray value u b of the foreground region is calculated according to the following formula:

[0063]

[0064] Further, in S4, the interference coherence coefficient image is segmented using the best threshold to obtain a water body region auxiliary extraction result, and the water body preliminary extraction result of the same month is fused with the water body region auxiliary extraction result to obtain a fused water body extraction result.

[0065] Further, in S5, the fused water body extraction result is subjected to morphological filtering processing, median filtering and connected domain rejection to obtain a final water body range.

[0066] The beneficial effects of the present application are: the weighted iterative threshold segmentation algorithm used in the present application can find a suitable segmentation threshold more accurately than the OTSU algorithm, and can exclude some weak object regions in the image, while using the interference coherence coefficient of long-time SAR image to weaken the influence of submerged plants and silt on the definition of water area edge; the present application can not only accurately calculate the segmentation threshold, but also avoid the influence of submerged plants and silt on the extraction result, thereby improving the accuracy of water body area extraction. BRIEF DESCRIPTION OF DRAWINGS

[0067] Figure 1 The flowchart of the water area extraction method for fusing radar backscattering and interference coherence coefficient;

[0068] Figure 2 for the SDWI water body index calculation result;

[0069] Figure 3 for the preliminary water body extraction result map of the embodiment of the application;

[0070] Figure 4 for the fused water body result map of the embodiment of the application;

[0071] Figure 5 for the morphological filtering result of the embodiment of the application;

[0072] Figure 6 for the final water body extraction result map of the embodiment of the application. DETAILED DESCRIPTION

[0073] The embodiments of the application are further described below with reference to the accompanying drawings.

[0074] As shown in the drawings, Figure 1 the application provides a water area extraction method fusing radar backscattering and interference coherence coefficient, comprising the following steps:

[0075] S1, acquiring a polarimetric SAR image, and pre-processing the polarimetric SAR image to obtain a backscattering coefficient image and an interference coherence coefficient image;

[0076] S2, calculating an SDWI water body index according to the backscattering coefficient images of different polarization modes;

[0077] S3, determining a preliminary water body extraction result by using an iterative threshold segmentation method according to the SDWI water body index;

[0078] S4, performing fusion processing on the preliminary water body extraction result based on the interference coherence coefficient image to generate a fused water body extraction result;

[0079] S5, performing refinement processing on the fused water body extraction result to obtain a final water body range.

[0080] In the embodiment of the application, in S1, the polarimetric SAR image is pre-processed by orbit correction, radiation correction, TOPS SAR image desmearing, cropping, multi-view, DN value conversion to dB, and geographic coding to obtain the backscattering coefficient image.

[0081] In S1, the polarimetric SAR image is pre-processed by orbit correction, radiation correction, TOPS SAR image segmentation, reverse geographic coding, TOPS SAR image desmearing, TOPS SAR image splicing, multi-view, interference, cropping, and geographic coding to obtain the interference coherence coefficient image.

[0082] In the embodiment of the present application, in S2, the SDWI water body index K SDWI The calculation formula is as follows:

[0083] K SDWI = ln (10*VV*VH) ;

[0084] In the formula, VV represents the backscattering coefficient in the vertical polarization mode, and VH represents the backscattering coefficient in the cross polarization mode.

[0085] In the embodiment of the present application, S3 comprises the following sub-steps:

[0086] S31, dividing the gray-scale image corresponding to the SAR image into a plurality of gray-scale levels, and calculating the probability of each gray-scale level;

[0087] S32, calculating the background class probability and the foreground class probability according to the probability of each gray-scale level;

[0088] Supposing that the segmentation threshold of the image is k, the gray-scale value less than the threshold k is divided into the background class a, and the gray-scale value greater than the threshold k is divided into the foreground class b.

[0089] S33, calculating the average gray-scale value of the background class and the average gray-scale value of the foreground class according to the background class probability and the foreground class probability;

[0090] S34, calculating the variance between different classes according to the background class probability, the foreground class probability, the average gray-scale value of the background class, and the average gray-scale value of the foreground class;

[0091] S35, calculating the variance between different elements in the background class and the variance between different elements in the foreground class according to the background class probability, the foreground class probability, the average gray-scale value of the background class, and the average gray-scale value of the foreground class;

[0092] S36, calculating the ratio of the sum of the inter-class variance and the intra-class variance according to the variance between different classes, the variance between different elements in the background class, and the variance between different elements in the foreground class;

[0093] S37, repeating S32-S36 to obtain the ratio of the sum of the inter-class variance and the intra-class variance of each gray-scale value, and extracting the gray-scale value corresponding to the maximum ratio;

[0094] S38, calculating the sum of the entropy of each gray-scale value, and extracting the gray-scale value corresponding to the maximum sum of the entropy;

[0095] S39, calculating the segmentation threshold according to the gray-scale value corresponding to the maximum ratio and the gray-scale value corresponding to the maximum sum of the entropy;

[0096] S310, determining the preliminary water body extraction result according to the segmentation threshold and the SDWI water body index.

[0097] In the embodiment of the present application, the calculation formula of the probability P i of the i-th gray scale in S31 is:

[0098]

[0099] wherein, n i represents the number of pixels with the gray value of i in the image, and N represents the total number of pixels of the gray image;

[0100] The calculation formula of the background class probability P a (k) in S32 is:

[0101]

[0102] wherein, k represents the segmentation threshold of the gray image;

[0103] The calculation formula of the foreground class probability P b (k) in S32 is:

[0104]

[0105] wherein, L represents the number of gray scales of the gray image;

[0106] The calculation formula of the average gray value m a (k) of the background class in S33 is:

[0107]

[0108] The calculation formula of the average gray value m b (k) of the foreground class in S33 is:

[0109]

[0110] The calculation formula of the variance σ0(k) 2 between different classes in S34 is:

[0111] σ0(k) 2 = P a (k)P b (k)(m a (k)-m b (k)) 2 ;

[0112] The calculation formula of the variance σ a between different elements in the background class in S35 is:

[0113]

[0114] The calculation formula of the variance σ bThe calculation formula of θ in S36 is:

[0115]

[0116] The calculation formula of θ in S36 is:

[0117]

[0118] The calculation formula of λ in S38 is:

[0119]

[0120] In the formula, ln(·) represents a logarithmic function;

[0121] The calculation formula of T1 in S39 is:

[0122]

[0123] In the formula, t a represents the gray value corresponding to the maximum ratio, and t b represents the gray value corresponding to the maximum entropy sum.

[0124] In the embodiment of the present application, S310 includes the following sub-steps:

[0125] S3101, dividing into a background region and a foreground region by using a division threshold value;

[0126] S3102, obtaining a weighted average gray value of the background region and a weighted average gray value of the foreground region by respectively weighting and averaging the background region and the foreground region based on the SDWI water body index;

[0127] S3103, taking a pixel greater than the weighted average gray value of the background region and less than the weighted average gray value of the foreground region as an uncertain region;

[0128] S3104, iteratively calculating the division threshold value of the uncertain region until the difference between the final division threshold value obtained and the division threshold value calculated in the last iteration is less than a convergence threshold value, and taking the final division threshold value as an optimal threshold value;

[0129] S3105, taking a region with a gray value lower than the optimal threshold value in the backscattering coefficient image as a preliminary water body result.

[0130] In the embodiment of the present application, in S3102, the calculation formula of the weighted average gray value u a of the background region is:

[0131]

[0132] In the formula, T1 represents a division threshold value, w(i) represents a weight of a pixel, f(x, y) is an SDWI water body index image, x represents a horizontal coordinate of each pixel on a two-dimensional remote sensing image, and y represents a vertical coordinate of each pixel on the two-dimensional remote sensing image.

[0133] In S3102, the weighted average gray value u of the foreground region is calculated. b The calculation formula of the weighted average gray value u of the foreground region is as follows:

[0134]

[0135] In the embodiment of the present application, in S4, the interference coherence coefficient image is segmented by using the optimal threshold value to obtain a water body region auxiliary extraction result, and the water body preliminary extraction result of the same month is fused with the water body region auxiliary extraction result to obtain a fused water body extraction result.

[0136] In the embodiment of the present application, in S5, the fused water body extraction result is subjected to morphological filtering processing, median filtering and connected domain rejection to obtain a final water body range.

[0137] The water body extraction result is subjected to a closing operation by using a disc-shaped structure element with a radius of 6 to smooth the boundary of the water body, so as to eliminate some small cavities generated after segmentation and connect the fine cracks, and make the image more complete and connected.

[0138] The present application will be described below in combination with specific embodiments.

[0139] In the embodiment of the present application, 226 periods of SLC data from 2016 to 2024 are used, and the SNAP10.0 software is used for processing. The Sentinel-1A SLC data is preprocessed to obtain a backscattering coefficient image and an interference coherence coefficient image. The calculated SDWI water body index result is as shown in Figure 2 The preliminary water body extraction result obtained by using the weighted iterative threshold segmentation algorithm on the SDWI water body index is as shown in Figure 3

[0140] The interference coherence coefficient image is subjected to threshold segmentation to obtain a water body auxiliary extraction result: the interference coherence coefficient image is subjected to threshold segmentation to weaken the interference of submerged plants and silt on the water area edge extraction, so as to cause the problem that the water body is not accurately extracted by using the OTSU algorithm combined with the SDWI. The interference coherence coefficient is a measure of the similarity between single-view complex images, that is, an index describing the difference in scattering characteristics of ground objects when taking pictures twice. Specifically, two periods of SAR images in the same month are taken as data to obtain the interference coherence coefficient image corresponding to the month, and the value of the interference coherence coefficient image less than 0.22 is taken as the water body.

[0141] ​The preliminary water extraction result and the auxiliary water extraction result are fused to obtain a water extraction result: the preliminary water extraction result using the backscattering coefficient and the auxiliary water extraction result using the radar interferometric coherence coefficient are fused to obtain a fused water extraction result. Influenced by the weak submerged plants and the silt, the echo signals thereof and the water echo signals are seriously mixed in the backscattering coefficient image, which leads to that the water cannot be completely extracted by using only the backscattering coefficient image. In order to solve this problem, the embodiment introduces the radar interferometric coherence coefficient information for supplement, performs threshold segmentation on the interferometric coherence coefficient image, extracts the missing part of the water, and fuses the water extraction result extracted by using the backscattering coefficient. The fused water extraction result is as shown in FIG. 6. Figure 4

[0142] The fused water extraction result is morphologically filtered: a disc-shaped structure element with a radius of 6 is used to perform a close operation on the water extraction result. The basic principle of the morphological processing is to collect information of the image by using the structure element, and to generate an output image by using the structure element to act on the input image. The morphological close operation can effectively fill small holes in the image and connect the adjacent broken edges, and can significantly eliminate the noise points caused by the inaccurate threshold segmentation, so as to optimize the fusion result of the water region. In this example, the segmented image is subjected to the close operation, the boundary of the water is smoothed, some small holes generated after the segmentation are eliminated, and the fine broken edges are connected, so that the image is more complete and connected, and the extraction accuracy of the water area is improved. The morphological filtering result is as shown in FIG. 7. Figure 5

[0143] The result after the morphological filtering is subjected to median filtering and connected domain rejection to obtain a final water distribution result: a 7x7 filtering window is selected to perform median filtering on the image after the morphological filtering, so as to eliminate the noise in the water extraction result. On this basis, the connected components with a pixel number less than 2000 are rejected by using the connected domain rejection, the water area with a small area is eliminated, and the main water area is retained. The final water result is as shown in FIG. 8. Figure 6

[0144] Those skilled in the art will appreciate that the embodiments described herein are presented to help the reader understand the principles of the present application, and should be understood as not limiting the protection scope of the present application to such specific statements and embodiments. Those skilled in the art can make various other specific modifications and combinations according to the technical inspirations disclosed in the present application without departing from the essence of the present application, and these modifications and combinations still fall within the protection scope of the present application.​​​

Claims

1. A method for water area extraction that integrates radar backscattering and interferometric coherence coefficients, characterized in that, Includes the following steps: S1. Acquire polarimetric SAR images and preprocess them to obtain backscattering coefficient images and interferometric coherence coefficient images; S2. Calculate the SDWI water index based on the backscattering coefficient images of different polarization modes; S3. Based on the SDWI water index, the iterative threshold segmentation method is used to determine the preliminary water extraction results; S4. Based on the interferometric coherence coefficient image, the preliminary water extraction results are fused to generate the fused water extraction results; S5. Refine the extracted water body results after fusion to obtain the final water body range; S3 includes the following sub-steps: S31. Divide the grayscale image corresponding to the SAR image into several grayscale levels and calculate the probability of each grayscale level. S32. Calculate the probability of the background class and the probability of the foreground class based on the probability of each gray level; S33. Calculate the average gray value of the background class and the average gray value of the foreground class based on the background class probability and the foreground class probability; S34. Calculate the variance between different classes based on the background class probability, the foreground class probability, the average gray value of the background class, and the average gray value of the foreground class. S35. Based on the background class probability, foreground class probability, average gray value of the background class, and average gray value of the foreground class, calculate the variance between different elements in the background class and the variance between different elements in the foreground class. S36. Based on the variances between different classes, the variances between different elements within the background class, and the variances between different elements within the foreground class, calculate the ratio of the sum of the variances between classes to the sum of the variances within classes. S37. Repeat S32-S36 to obtain the ratio of the sum of the inter-class variance and the intra-class variance for each gray value, and extract the gray value corresponding to the largest ratio. S38. Calculate the sum of the entropy of each gray value and extract the gray value corresponding to the maximum sum of entropy; S39. Calculate the segmentation threshold based on the gray value corresponding to the maximum ratio and the gray value corresponding to the sum of the maximum entropies; S310. Based on the classification threshold and SDWI water index, determine the preliminary water extraction results.

2. The water area extraction method based on the fusion of radar backscattering and interferometric coherence coefficients according to claim 1, characterized in that, In S1, the polarimetric SAR image is preprocessed by orbit correction, radiometric correction, TOPS SAR image deskewing, cropping, multi-look, DN value to dB conversion and geocoding to obtain the backscattering coefficient image. In step S1, the polarimetric SAR image is preprocessed through orbit correction, radiometric correction, TOPS SAR image segmentation, inverse geocoding, TOPS SAR image deskewing, TOPS SAR image stitching, multi-view, interferometry, cropping, and geocoding to obtain an interferometric coherence coefficient image.

3. The water area extraction method based on the fusion of radar backscattering and interferometric coherence coefficients according to claim 1, characterized in that, In S2, the SDWI water index K SDWI The calculation formula is: ； In the formula, VV This represents the backscattering coefficient for vertical polarization. VH This represents the backscattering coefficient for cross-polarization.

4. The water area extraction method based on the fusion of radar backscattering and interferometric coherence coefficients according to claim 1, characterized in that, In S31, the first i The probability of each gray level P i The calculation formula is: ； In the formula, n i Indicates the grayscale value in the image i The number of pixels, N Represents the total number of pixels in a grayscale image; In S32, the probability of the background class P a ( k The formula for calculating ) is: ； In the formula, k The segmentation threshold for a grayscale image; In S32, the foreground class probability P b ( k The calculation formula for ); ； In the formula, L Indicates the number of gray levels in a grayscale image; In S33, the average gray value of the background class m a ( k The formula for calculating ) is: ； In S33, the average gray value of the foreground class m b ( k The formula for calculating ) is: ； In S34, the variance between different classes The calculation formula is: ； In S35, the variance between different elements within the background class σ a The calculation formula is: ； In S35, the variance between different elements within the foreground class σ b The calculation formula is: ； In S36, the ratio of the sum of the between-class variances to the sum of the within-class variances. θ The calculation formula is: ； In S38, the sum of the entropy of the gray values λ The calculation formula is: ； In the formula, ln (·) represents a logarithmic function; In S39, the segmentation threshold T The formula for calculating 1 is: ； In the formula, t a This represents the grayscale value corresponding to the maximum ratio. t b This represents the grayscale value corresponding to the sum of maximum entropy.

5. The method for extracting water bodies by fusing radar backscattering and interferometric coherence coefficients according to claim 1, characterized in that, S310 includes the following sub-steps: S3101. Using the segmentation threshold, the region is divided into a background region and a foreground region; S3102. Based on the SDWI water index, the weighted average values ​​of the background and foreground regions are calculated separately to obtain the weighted average gray value of the background region and the weighted average gray value of the foreground region. S3103. Pixels with a weighted average gray value greater than that of the background region and less than that of the foreground region are considered as uncertain regions. S3104. Iteratively calculate the threshold for dividing the uncertain region until the difference between the final threshold and the threshold calculated in the previous iteration is less than the convergence threshold, and take the final threshold as the optimal threshold. S3105. The region in the backscattering coefficient image with a gray value lower than the optimal threshold is taken as the preliminary water body result.

6. The method for extracting water bodies by fusing radar backscattering and interferometric coherence coefficients according to claim 5, characterized in that, In step S3102, the weighted average gray value of the background area u a The calculation formula is: ； In the formula, T 1 represents the threshold for segmentation. w ( i ) represents the pixel weight. To represent SDWI water index images, x This represents the x-coordinate of each pixel in the two-dimensional remote sensing image. y This represents the vertical coordinate of each pixel in the two-dimensional remote sensing image. In step S3102, the weighted average gray value of the foreground region u b The calculation formula is: 。 7. The method for extracting water bodies by fusing radar backscattering and interferometric coherence coefficients according to claim 1, characterized in that, In step S4, the interference coherence coefficient image is segmented using the optimal threshold to obtain the water body region auxiliary extraction result. The preliminary water body extraction result of the same month is then fused with the water body region auxiliary extraction result to obtain the fused water body extraction result.

8. The method for extracting water bodies by fusing radar backscattering and interferometric coherence coefficients according to claim 1, characterized in that, In step S5, the fused water body extraction results are subjected to morphological filtering, median filtering, and connected component removal to obtain the final water body range.

Images