Water body identification and extraction method and system based on GF6-WFV data based on bilateral red band

Abstract

This paper proposes a water body identification and extraction method and system based on bilateral red band GF6-WFV data, including: preprocessing GF6 satellite WFV image data to obtain atmospherically corrected multi-angle observation data for each satellite band; calculating the reflectivity of each satellite band, normalized to the same zenith angle after atmospheric correction, to obtain the angularly normalized reflectivity of each band; superimposing bands with high and low reflectivity effects and then normalizing them to eliminate the noise impact of areas including shadows and dark surfaces; calculating a targeted high-resolution water index and the normalized difference vegetation index (NDVI). This paper reveals the differences in sensitivity of different bands in water body monitoring and can effectively distinguish water bodies from other land cover types in complex natural and urban environments, making the water body identification process more accurate.

Description

Technical Field

[0001] The present invention belongs to the field of spatial information technology, and in particular relates to a method and system for identifying and extracting water bodies from GF6-WFV data based on a bilateral red band. Background Art

[0002] With the widespread development of remote sensing technology in lake research and applications, efficiently extracting useful information from massive amounts of data has become a core issue in remote sensing applications. Currently, the main methods for extracting water body information using remote sensing technology can be categorized into two main categories: single-band and multi-band methods. Single-band methods primarily identify water body information by processing and analyzing remote sensing imagery captured in a single band, based on the spectral characteristics of water bodies in specific wavelengths. Specifically, water strongly absorbs near-infrared light, while vegetation and dry soil exhibit strong reflectivity in this wavelength band. In practice, selecting the near-infrared band in remote sensing images and combining it with appropriate thresholding can effectively separate water bodies from the image. This method is simple to operate and computationally efficient, making it suitable for areas with high water contrast and minimal background interference. However, in complex scenes, such as those with blurred water edges or mixed pixels, the extraction accuracy of single-band methods may be affected to some extent.

[0003] In contrast, multi-band methods leverage information from multiple bands in multispectral imagery to enhance and extract water bodies through spectral combination or index algorithms (such as the Normalized Difference Water Index (NDWI)). Multi-band methods are better able to handle situations where water bodies are mixed with the background, offering advantages in accurate extraction and classification. Compared with single-band thresholding methods, multi-band methods more comprehensively utilize spectral information, significantly improving water identification accuracy. Their core concept is to extract water-related information from multiple bands and construct a combination formula, such as a spectral index or ratio, using specific algorithms to enhance the water signal while suppressing background interference. Water index methods are particularly popular, with typical water indices including the Normalized Difference Water Index (NDWI) and the Modified Normalized Difference Water Index (MNDWI). NDWI typically utilizes the difference between the green and near-infrared bands to highlight water signals, while MNDWI further incorporates the mid-infrared band to better suppress interference from background objects such as buildings and bare land. These methods demonstrate strong robustness and adaptability when extracting water body information, particularly in complex surface environments, significantly improving the accuracy and refinement of water body identification. Based on the spectral relationship between bands, a new water index (NWI) model was proposed and effectively validated in typical water bodies such as reservoirs, lakes, and rivers. Furthermore, through studies of rivers and lakes in diverse regions under varying climatic conditions and seasonal variations, a multi-band water index (MBWI) model was further developed to enhance the applicability and accuracy of water body information extraction.

[0004] The Gaofen-6 satellite, officially commissioned on March 21, 2019, is a key component of my country's low-orbit, high-resolution optical remote sensing satellite fleet. It features high resolution, wide coverage, high-quality, and efficient imaging. The satellite's multispectral medium-resolution wide-swath camera (WFV) has an observation swath exceeding 800 kilometers, a spatial resolution of 16 meters, and covers eight spectral bands. The GF6-WFV camera boasts high spectral sensitivity within each band, enabling it to capture subtle spectral variations in surface features, offering significant advantages in vegetation classification and land cover change.

[0005] However, current water body information extraction methods often rely on simple spectral threshold judgments or pixel-based classification models, and generally lack specialized optimization and design for the spectral characteristics of the unique bands of GF6-WFV images. When faced with different geographical environments, seasonal changes, or diverse water body characteristics, there is a high misidentification rate and low efficiency. Especially in the boundary areas where urban and natural environments alternate, it is often impossible to accurately distinguish water bodies from other surface covers. In addition, traditional water body extraction technologies mainly focus on natural water bodies such as lakes, rivers, and wetlands, but they often face identification difficulties for relatively small and diverse aquaculture waters such as fish ponds, shrimp ponds, and salt pans. These aquaculture waters are usually scattered and irregular in size, and are easily affected by spectral interference from surrounding vegetation, soil, buildings, and other objects, making it difficult to ensure extraction accuracy. Summary of the Invention

[0006] The present invention proposes a water body identification and extraction method and system based on the bilateral red band of GF6-WFV data. By deeply analyzing the reflection and absorption characteristics of each band in GF6-WFV imagery on water bodies, the sensitivity differences of different bands in water body monitoring are revealed. In complex natural and urban environments, it can effectively distinguish water bodies from other surface cover types, making the water body identification process more accurate.

[0007] To achieve the above object, the technical solution of the present invention is achieved as follows:

[0008] A water body identification and extraction method based on the bilateral red band of GF6-WFV data includes:

[0009] Preprocessing of WFV image data from S1 and GF6 satellites to obtain atmospherically corrected multi-angle observation data for each band of the satellite;

[0010] S2. Based on the atmospherically corrected multi-angle observation data of each satellite band, the reflectivity of each satellite band is calculated to the same zenith angle to obtain the reflectivity of each band after angle normalization.

[0011] S3. Based on the normalized reflectivity of each band, the bands with high reflectivity and low reflectivity are superimposed and normalized to eliminate the noise impact of shadows and dark surfaces, and a targeted high-resolution water index is calculated.

[0012] S4. Calculate the Normalized Difference Vegetation Index (NDVI) based on the normalized reflectance of each band angle. Areas with NDVI greater than 0 are fragmented water bodies, and areas with NDVI less than 0 are large natural water bodies.

[0013] Furthermore, step S1 includes:

[0014] S101, based on the original image data acquired by the Gaofen-6 satellite GF6, extract all XML files related to the sun angle, and parse the solar zenith angle and solar azimuth angle;

[0015] S102: Select the minimum, middle, and maximum solar zenith angles and their corresponding solar azimuth angles to obtain the corresponding satellite zenith angles and satellite azimuth angles;

[0016] S103. For each zenith angle, traverse different aerosol optical depths (AODs), and generate corresponding 6S model input files for each AOD value and different satellite bands. Perform atmospheric correction using the 6S model and output the true surface reflectance at different angles and bands.

[0017] Furthermore, step S2 includes:

[0018] The calculation is performed using the bidirectional reflectance distribution function that describes the bidirectional reflectance characteristics of the earth's surface;

[0019] ;

[0020] Where, is the solar zenith angle , satellite zenith angle At, relative azimuth Bidirectional reflectance distribution function. Relative azimuth is the difference between the sun's azimuth and the satellite's azimuth, is the volume scattering kernel, is the geometric optics kernel, both of which are functions of three angles. Indicates bands, model parameters 、 、 Respectively expressed in The proportion of the band isotropic component, volume scattering component and geometric optics scattering component in the reflectivity;

[0021] The reflectivity of each band after angle normalization is calculated.

[0022] Furthermore, step S3 includes:

[0023] ;

[0024] Wherein, TGFWI represents the targeted high-resolution water index, B2 represents the green band reflectance calculated in step S2, B4 represents the near-infrared band reflectance calculated in step S2, B5 represents the red edge 1 band reflectance calculated in step S2, and B6 represents the red edge 2 band reflectance calculated in step S2.

[0025] Furthermore, step S4 includes:

[0026] ;

[0027] Wherein, NDVI represents the Normalized Difference Vegetation Index, B3 represents the red light band reflectance calculated in step S2, and B4 represents the near infrared band reflectance calculated in step S2.

[0028] On the other hand, the present invention also proposes a water body identification and extraction system based on the bilateral red band of GF6-WFV data, comprising:

[0029] Preprocessing module: Preprocessing of GF6 satellite WFV image data to obtain atmospherically corrected multi-angle observation data of each satellite band;

[0030] Reflectivity module: Based on the atmospherically corrected multi-angle observation data of each satellite band, the reflectivity of each satellite band is calculated to the same zenith angle, and the reflectivity of each band is obtained after angle normalization.

[0031] Water Index Module: Based on the normalized reflectivity of each band angle, the bands with high reflectivity and low reflectivity are superimposed and normalized separately to eliminate the noise impact of areas including shadows and dark surfaces, and calculate a targeted high-resolution water index.

[0032] Vegetation Index Module: Calculates the Normalized Vegetation Index (NDVI) based on the normalized reflectance of each band angle. Areas with NDVI greater than 0 are fragmented water bodies, and areas with NDVI less than 0 are large natural water bodies.

[0033] Furthermore, the preprocessing module includes:

[0034] Data unit: Based on the original image data from the Gaofen-6 satellite GF6, all XML files related to the sun angle are extracted and the solar zenith angle and solar azimuth angle are parsed;

[0035] Selection unit: select the minimum, middle and maximum solar zenith angles and their corresponding solar azimuth angles, and obtain the corresponding satellite zenith angles and satellite azimuth angles;

[0036] Correction unit: For each zenith angle, it traverses different aerosol optical depths (AODs) and generates corresponding 6S model input files for each AOD value and different satellite bands. It performs atmospheric correction through the 6S model and outputs the true surface reflectance at different angles and bands.

[0037] Furthermore, the reflectivity module includes:

[0038] The calculation is performed using the bidirectional reflectance distribution function that describes the bidirectional reflectance characteristics of the earth's surface;

[0039] ;

[0040] Where, is the solar zenith angle , satellite zenith angle At, relative azimuth Bidirectional reflectance distribution function. Relative azimuth is the difference between the sun's azimuth and the satellite's azimuth, is the volume scattering kernel, is the geometric optics kernel, both of which are functions of three angles. Indicates bands, model parameters 、 、 Respectively expressed in The proportion of the band isotropic component, volume scattering component and geometric optics scattering component in the reflectivity;

[0041] The reflectivity of each band after angle normalization is calculated.

[0042] Furthermore, the water index module includes:

[0043] ;

[0044] Wherein, TGFWI represents the targeted high-resolution water index, B2 represents the green band reflectance calculated in step S2, B4 represents the near-infrared band reflectance calculated in step S2, B5 represents the red edge 1 band reflectance calculated in step S2, and B6 represents the red edge 2 band reflectance calculated in step S2.

[0045] Furthermore, the vegetation index module includes:

[0046] ;

[0047] Wherein, NDVI represents the Normalized Difference Vegetation Index, B3 represents the red light band reflectance calculated in step S2, and B4 represents the near infrared band reflectance calculated in step S2.

[0048] Compared with the prior art, the present invention has the following beneficial effects:

[0049] 1. Through the solution of the present invention, not only can high-precision spatial positioning of water body identification and extraction be achieved, but also dynamic monitoring can be carried out, and water body changes and water quality status can be tracked in real time. In particular, in areas with complex water types and rapid environmental changes, such as urban rivers, lakes and wetlands, it is possible to achieve refined classification and change analysis of water bodies, providing more accurate data support. The present invention not only optimizes the limitations of traditional water body identification methods, but also further improves the accuracy and efficiency of water body identification, solving complex environmental problems that were previously difficult to deal with based on single spectral data or rough classification methods.

[0050] 2. The present invention utilizes high-resolution data of multiple bands in the GF-6 satellite WFV imagery, combined with the unique spectral characteristics of water bodies, to achieve high-precision water body identification under a wider range of environmental conditions. Specifically, the present invention extracts the unique spectral characteristics of water bodies and effectively distinguishes water bodies from other surface cover types, such as vegetation, soil, buildings, etc., through systematic and in-depth analysis and modeling of the reflectivity differences in each band, thereby significantly reducing the impact of interference sources on water body identification and greatly improving the accuracy and robustness of identification. The differences in spectral characteristics between aquaculture waters and natural water bodies are deeply analyzed, and the precise capture of water body boundaries and morphology with high-resolution data is combined to achieve effective extraction of aquaculture waters. Especially in a diverse environmental context, the present invention can automatically identify subtle differences between aquaculture waters and surrounding land features, thereby avoiding misjudgment and improving the accuracy and reliability of extraction. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 It is a flow chart of Example 1 of the present invention.

[0052] Figure 2 This is a GF6-WFV image data diagram of Example 1 of the present invention.

[0053] Figure 3 3 is a comparison chart of the water index extraction results of Example 1 of the present invention.

[0054] Figure 3 (a) is the normalized difference water index NDWI extraction result diagram, and (b) is the targeted high-resolution water index TGFWI extraction result diagram of the present invention.

[0055] Figure 4 This is an NDVI water body classification effect diagram of Example 1 of the present invention.

[0056] Figure 4 (a) is a fragmented water body map, and (b) is a large-scale natural water body map. DETAILED DESCRIPTION

[0057] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments may be combined with each other.

[0058] In order to make the purpose and features of the present invention more obvious and easy to understand, the present invention is further described below with reference to the accompanying drawings and specific embodiments.

[0059] Example 1:

[0060] This invention relates to a method for identifying and extracting water bodies from GF6-WFV data based on the bilateral red band. The core of this technical solution is to acquire multi-angle, multi-temporal remote sensing data from the Gaofen-6 satellite and, in combination with existing preprocessing techniques, perform targeted index calculations on the bilateral red band, ultimately achieving efficient and accurate water body identification and extraction. The method was implemented on a computer equipped with an Intel second-generation Core i3-2100 @ 3.10GHz dual-core processor, 4GB of memory, and an Nvidia GeForce GT 630 graphics card.

[0061] The WFV image data of S1 and GF6 satellites are preprocessed to obtain multi-angle observation data of each satellite band after atmospheric correction.

[0062] This step is part of the GF6 satellite image data preprocessing phase. Based on the raw WFV image data acquired by the Gaofen-6 satellite, GF6 also provides a satellite navigation file with a .nav suffix. These files contain navigation parameters (orbit, attitude, etc.) and an XML file related to solar angles. In this example, IDL code is written to read the .nav file, extract the XML file related to solar angles, and parse it to determine the solar zenith angle (SolarZenith) and solar azimuth angle (SolarAzimuth).

[0063] Specifically, it involves traversing all XML files and extracting the solar zenith angle and azimuth information. All zenith angle and azimuth values ​​are stored in the sz and sa arrays, and sorted according to these values. Finally, the minimum, middle, and maximum zenith angles (sz_s) and their corresponding solar azimuth angles (sa_s) are selected as inputs to the lookup table. The lookup table refers to the lookup table for 6S atmospheric correction, and the atmospheric correction coefficient x is output through the lookup table. a 、x b 、x c , and then calculate the surface reflectivity.

[0064] For each selected solar zenith angle (sz_s[t]), extract the corresponding satellite navigation file (.nav) data to obtain the satellite zenith angle (satZenith) and satellite azimuth angle (satAzimuth). For each zenith angle, traverse different aerosol optical depths (AODs) and generate corresponding 6S input files for each AOD value (from 0 to 0.5, with a step size of 0.1) and 8 different bands (defined in GF6-WFV). Run the 6S program and use the 6S model to calculate the atmospheric correction coefficient x. a 、x b 、x c Perform atmospheric correction and output the true surface reflectivity at different angles and bands.

[0065] S2: Based on the atmospherically corrected multi-angle observation data of each satellite band, the reflectivity of each satellite band is calculated by unifying them to the same zenith angle to obtain the angle-normalized reflectivity of each band.

[0066] The multi-angle observation data in this step is the true surface reflectivity after atmospheric correction as described in step 1. The reflectivity of a single position varies due to different angles.

[0067] This step uses a kernel-driven model to describe the bidirectional reflectance characteristics of the Earth's surface using a linear combination of physically meaningful kernels. The model consists of three scattering components (isotropic, volume scattering, and geometric optics scattering) that are linearly weighted and vary with angle and illumination. The specific formula (1) is as follows:

[0068] ; (1)

[0069] Where, is the solar zenith angle , satellite zenith angle At, relative azimuth Bidirectional reflectance distribution function. Relative azimuth is the difference between the sun's azimuth and the satellite's azimuth.

[0070] is the volume scattering kernel, is the geometric optics kernel, both of which are functions of three angles. Indicates bands, model parameters 、 、 Respectively expressed in The proportion of the band isotropic component, volume scattering component and geometric optical scattering component in the reflectivity. Under the conditions of known band and angle, the corresponding BRDF prototype is found according to the designed angle normalization lookup table to obtain the corresponding model parameters 、 、 , combined with the above kernel-driven model to calculate the reflectivity and zenith reflectivity, complete the reflectivity angle normalization, and eliminate the influence of surface anisotropy.

[0071] S3. Based on the normalized reflectivity of each band angle, the bands with high reflection effects and the bands with low reflection effects are superimposed and normalized respectively to eliminate the noise influence of areas including shadows and dark surfaces, and calculate the targeted high-resolution water index.

[0072] In this step, the Targeted GaoFen Water Index (TGFWI) is calculated based on the normalized reflectance of B1 (blue band reflectance), B2 (green band reflectance), B3 (red band reflectance), B4 (near-infrared band reflectance), B5 (red edge 1 band reflectance), B6 ​​(red edge 2 band reflectance), B7 (purple edge band reflectance), and B8 (yellow edge band reflectance) obtained in step S2. By observing the spectral response function and statistically comparing the reflectances of different bands, it is found that B2 and B5 have a high-reflection effect, while B4 and B6 have a low-reflection effect. By superimposing the two to enhance the corresponding effects, (B2 + B5) and (B4 + B6) can be regarded as high-reflection and low-reflection overall, respectively, and then normalized.

[0073] Because the reflection of the red edge 1 band is stronger than that of the near-infrared and red edge 2 bands, the noise influence of areas including shadows and dark surfaces is eliminated, and the focus is on extracting scattered water bodies.

[0074] In order to establish the high reflectivity bands B2 and B5 for GF6-WFV to enhance its high reflectivity effect, and the low reflectivity bands B4 and B6 to enhance the low reflectivity effect of water bodies, and then borrow the unique characteristics of the water spectrum, that is, the reflection of the red edge 1 is stronger than the reflection of the near-infrared band and the red edge 2 band, the water body index TGFWI is calculated using formula (2):

[0075] ; (2)

[0076] The TGFWI water index is designed to eliminate non-water pixels and can extract water bodies from GF satellite images. It is not only suitable for the efficient extraction of large-scale natural water bodies, but can also accurately identify and extract surrounding aquaculture waters, demonstrating wide adaptability and diverse application capabilities. The index can also suppress dark building surfaces in urban background areas and extract urban water bodies.

[0077] S4. Calculate the Normalized Difference Vegetation Index (NDVI) based on the normalized reflectance of each band angle. Areas with NDVI greater than 0 are fragmented water bodies, and areas with NDVI less than 0 are large natural water bodies.

[0078] In this step, the Normalized Difference Vegetation Index (NDVI) is calculated and a threshold is set to distinguish large water bodies from scattered water bodies. The NDVI is calculated using the following formula (3), where B3 is the reflectance of the red light band and B4 is the reflectance of the near-infrared band:

[0079] ; (3)

[0080] Select the GF6-WFV image data of a local area, the detailed location is as follows Figure 2 As shown, the area includes three lakes.

[0081] The water bodies in the area were extracted using NDWI and the TGFWI method of the present invention on ENVI. The extraction results of each water body index are as follows: Figure 3 shown. Figure 3 (a) and (b) are the extraction results of the normalized difference water index NDWI and the targeted high-resolution water index TGFWI of the GF6-WFV image, respectively.

[0082] Depend on Figure 3 It can be seen that all kinds of water index can enhance the water quality, but there are obvious differences and different effects. The normalized difference water index NDWI ( Figure 3 (a) The water body is significantly enhanced, and the lake water and background can be separated, but some of the lake water area has not been enhanced. Targeted high-resolution water body index TGFWI ( Figure 3 (b) can effectively separate the water features from the background and almost completely enhance the lake area. To further distinguish large areas of water from fragmented waters, the Normalized Difference Vegetation Index (NDVI) is used for classification. Figure 4 The extraction effect can be seen, where the area with NDVI greater than 0 is a fragmented water body ( Figure 4 (a)), mainly aquaculture, and the area with NDVI less than 0 is a large area of ​​natural water bodies ( Figure 4 (b)), mainly naturally formed lakes and rivers.

[0083] Example 2:

[0084] This embodiment proposes a water body identification and extraction system based on the bilateral red band of GF6-WFV data, including:

[0085] Preprocessing module: Preprocessing of GF6 satellite WFV image data to obtain atmospherically corrected multi-angle observation data of each satellite band;

[0086] Reflectivity module: Based on the atmospherically corrected multi-angle observation data of each satellite band, the reflectivity of each satellite band is calculated to the same zenith angle, and the reflectivity of each band is obtained after angle normalization.

[0087] Water Index Module: Based on the normalized reflectivity of each band angle, the bands with high reflectivity and low reflectivity are superimposed and normalized separately to eliminate the noise impact of areas including shadows and dark surfaces, and calculate a targeted high-resolution water index.

[0088] Vegetation Index Module: Calculates the Normalized Vegetation Index (NDVI) based on the normalized reflectance of each band angle. Areas with NDVI greater than 0 are fragmented water bodies, and areas with NDVI less than 0 are large natural water bodies.

[0089] Among them, the preprocessing module includes:

[0090] Data unit: Based on the original image data from the Gaofen-6 satellite GF6, all XML files related to the sun angle are extracted and the solar zenith angle and solar azimuth angle are parsed;

[0091] Selection unit: select the minimum, middle and maximum solar zenith angles and their corresponding solar azimuth angles, and obtain the corresponding satellite zenith angles and satellite azimuth angles;

[0092] Correction unit: For each zenith angle, it traverses different aerosol optical depths (AODs) and generates corresponding 6S model input files for each AOD value and different satellite bands. It performs atmospheric correction through the 6S model and outputs the true surface reflectance at different angles and bands.

[0093] The reflectivity module includes:

[0094] The calculation is performed using the bidirectional reflectance distribution function that describes the bidirectional reflectance characteristics of the earth's surface;

[0095] ;

[0096] Where, is the solar zenith angle , satellite zenith angle At, relative azimuth Bidirectional reflectance distribution function. Relative azimuth is the difference between the sun's azimuth and the satellite's azimuth, is the volume scattering kernel, is the geometric optics kernel, both of which are functions of three angles. Indicates bands, model parameters 、 、 Respectively expressed in The proportion of the band isotropic component, volume scattering component and geometric optics scattering component in the reflectivity;

[0097] The reflectivity of each band after angle normalization is calculated.

[0098] The water index module includes:

[0099] ;

[0100] Wherein, TGFWI represents the targeted high-resolution water index, B2 represents the green band reflectance calculated in step S2, B4 represents the near-infrared band reflectance calculated in step S2, B5 represents the red edge 1 band reflectance calculated in step S2, and B6 represents the red edge 2 band reflectance calculated in step S2.

[0101] The vegetation index module includes:

[0102] ;

[0103] Wherein, NDVI represents the Normalized Difference Vegetation Index, B3 represents the red light band reflectance calculated in step S2, and B4 represents the near infrared band reflectance calculated in step S2.

[0104] The GF6-WFV data water body identification and extraction system based on the bilateral red band proposed in this embodiment can implement the GF6-WFV data water body identification and extraction method based on the bilateral red band described in Example 1, and has the same technical effect as Example 1.

[0105] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A water body identification and extraction method based on GF6-WFV data with bilateral red band, characterized in that: include: Preprocessing of WFV image data from S1 and GF6 satellites to obtain atmospherically corrected multi-angle observation data for each band of the satellite; S2. Based on the atmospherically corrected multi-angle observation data of each satellite band, the reflectivity of each satellite band is calculated to the same zenith angle to obtain the reflectivity of each band after angle normalization. S3. Based on the normalized reflectivity of each band, the bands with high reflectivity and low reflectivity are superimposed and normalized to eliminate the noise impact of shadows and dark surfaces, and a targeted high-resolution water index is calculated. S4. Calculate the Normalized Difference Vegetation Index (NDVI) based on the normalized reflectance of each band angle. Areas with NDVI greater than 0 are fragmented water bodies, and areas with NDVI less than 0 are large natural water bodies. Step S3 includes: ； Wherein, TGFWI represents the targeted high-resolution water index, B2 represents the green band reflectance calculated in step S2, B4 represents the near-infrared band reflectance calculated in step S2, B5 represents the red edge 1 band reflectance calculated in step S2, and B6 represents the red edge 2 band reflectance calculated in step S2; Step S4 includes: ； Wherein, NDVI represents the Normalized Difference Vegetation Index, B3 represents the red light band reflectance calculated in step S2, and B4 represents the near infrared band reflectance calculated in step S2.

2. The water body identification and extraction method based on the bilateral red band of GF6-WFV data according to claim 1 is characterized in that: Step S1 includes: S101, based on the original image data acquired by the Gaofen-6 satellite GF6, extract all XML files related to the sun angle, and parse the solar zenith angle and solar azimuth angle; S102: Select the minimum, middle, and maximum solar zenith angles and their corresponding solar azimuth angles to obtain the corresponding satellite zenith angles and satellite azimuth angles; S103. For each zenith angle, traverse different aerosol optical depths (AODs), and generate corresponding 6S model input files for each AOD value and different satellite bands. Perform atmospheric correction using the 6S model and output the true surface reflectance at different angles and bands.

3. The water body identification and extraction method based on the bilateral red band of GF6-WFV data according to claim 1 is characterized in that: Step S2 includes: The calculation is performed using the bidirectional reflectance distribution function that describes the bidirectional reflectance characteristics of the earth's surface; ； Where, is the solar zenith angle , satellite zenith angle At, relative azimuth Bidirectional reflectance distribution function; relative azimuth is the difference between the sun's azimuth and the satellite's azimuth, is the volume scattering kernel, is the geometric optics kernel, both of which are functions of three angles; Indicates bands, model parameters 、 、 Respectively expressed in The proportion of the band isotropic component, volume scattering component and geometric optics scattering component in the reflectivity; The reflectivity of each band after angle normalization is calculated.

4. A water body identification and extraction system based on GF6-WFV data with bilateral red band, characterized by: include: Preprocessing module: Preprocessing of GF6 satellite WFV image data to obtain atmospherically corrected multi-angle observation data of each satellite band; Reflectivity module: Based on the atmospherically corrected multi-angle observation data of each satellite band, the reflectivity of each satellite band is calculated to the same zenith angle, and the reflectivity of each band is obtained after angle normalization. Water Index Module: Based on the normalized reflectivity of each band angle, the bands with high reflectivity and low reflectivity are superimposed and normalized separately to eliminate the noise impact of areas including shadows and dark surfaces, and calculate a targeted high-resolution water index. Vegetation Index Module: Calculates the Normalized Vegetation Index (NDVI) based on the reflectance normalized by each band angle. Areas with NDVI greater than 0 are fragmented water bodies, and areas with NDVI less than 0 are large areas of natural water bodies. The water index module includes: ； Wherein, TGFWI represents the targeted high-resolution water index, B2 represents the green band reflectance calculated in step S2, B4 represents the near-infrared band reflectance calculated in step S2, B5 represents the red edge 1 band reflectance calculated in step S2, and B6 represents the red edge 2 band reflectance calculated in step S2; The vegetation index module includes: ； Wherein, NDVI represents the Normalized Difference Vegetation Index, B3 represents the red light band reflectance calculated in step S2, and B4 represents the near infrared band reflectance calculated in step S2.

5. The GF6-WFV data water body identification and extraction system based on the bilateral red band according to claim 4 is characterized in that: The preprocessing modules include: Data unit: Based on the original image data from the Gaofen-6 satellite GF6, all XML files related to the sun angle are extracted and the solar zenith angle and solar azimuth angle are parsed; Selection unit: select the minimum, middle and maximum solar zenith angles and their corresponding solar azimuth angles, and obtain the corresponding satellite zenith angles and satellite azimuth angles; Correction unit: For each zenith angle, it traverses different aerosol optical depths (AODs) and generates corresponding 6S model input files for each AOD value and different satellite bands. It performs atmospheric correction through the 6S model and outputs the true surface reflectance at different angles and bands.

6. The GF6-WFV data water body identification and extraction system based on the bilateral red band according to claim 4 is characterized in that: The reflectivity module includes: The calculation is performed using the bidirectional reflectance distribution function that describes the bidirectional reflectance characteristics of the earth's surface; ； Where, is the solar zenith angle , satellite zenith angle At, relative azimuth Bidirectional reflectance distribution function; relative azimuth is the difference between the sun's azimuth and the satellite's azimuth, is the volume scattering kernel, is the geometric optics kernel, both of which are functions of three angles; Indicates bands, model parameters 、 、 Respectively expressed in The proportion of the band isotropic component, volume scattering component and geometric optics scattering component in the reflectivity; The reflectivity of each band after angle normalization is calculated.

Images