Trans-climate zone surface water resource active and passive cooperative remote sensing extraction method

Abstract

The invention relates to a cross-climate zone surface water resource active and passive cooperative remote sensing extraction method, which comprises the following steps of: preprocessing by taking a satellite historical image as a basis to obtain an optical synthetic satellite image and a radar synthetic satellite image of a target research area, and training to obtain an optical image water body recognition model; the method comprises the following steps: acquiring a target research area, connecting an optimal water body classification threshold value under a radar synthetic satellite image, respectively realizing water body distribution identification in two aspects of model identification and classification threshold value, and finally fusing identification results of the two modes to obtain water body distribution of the target research area corresponding to each historical acquisition moment; according to the design scheme, optical data and radar data are deeply fused, large-scale and long-time-sequence rapid automatic water body mapping is achieved, the precision and reliability of an extraction result are remarkably improved, the limitation of a single data source can be made up, important support can be provided for dynamic monitoring and scientific management of drainage basin water resources, and the method is suitable for popularization and application. The method is especially suitable for water resource monitoring of cross-climate zone complex drainage basins.

Description

Technical Field

[0001] The invention relates to an active and passive coordinated remote sensing extraction method for surface water resources across climate zones, belonging to the technical field of surface water body detection. Background Art

[0002] Surface water resources are an important part of the global ecosystem, with significant environmental regulation functions and ecological benefits, and are closely related to human health, living environment, agricultural production, and economic activities. In the Nile River Basin in Africa, an important region spanning tropical, subtropical, and arid climate zones, surface water resources are not only a key factor in maintaining ecological balance and social development, but also an important basis for the coordinated development of countries in the basin. With the intensification of global climate change and human activities, the distribution and dynamic changes of water resources in the basin have received increasing attention.

[0003] Effectively detecting the temporal and spatial distribution characteristics of surface water bodies and dynamically monitoring their changing patterns are of great significance to complex regions across climate zones such as the Nile River Basin. On the one hand, such monitoring can provide data support for the scientific management of water resources and help stabilize and restore regional ecosystems; on the other hand, it can provide a scientific basis for climate change impact assessments and transnational water resource management strategies.

[0004] Remote sensing technology has been widely used in the extraction and monitoring of surface water resources due to its advantages of real-time, periodic, large-scale coverage and multi-source data fusion. Optical remote sensing technology, such as Sentinel-2 images, has excellent performance in low, medium and high resolution surface water extraction with rich spectral information, but it is easily affected by interference environments such as cloud shadows, complex terrain and snow. Radar remote sensing technology, such as Sentinel-1 images, can penetrate clouds and have all-weather imaging capabilities, which are suitable for surface water extraction under complex weather conditions. However, radar images are greatly interfered by speckle noise, and the image grayscale is not as continuous as optical images. At the same time, the scattering characteristics of buildings and mountain shadows on radar data will also cause certain misjudgments.

[0005] In recent years, the method of active and passive remote sensing data collaboration has gradually received attention. Some researchers have tried to combine the complementary advantages of optical and radar remote sensing data to improve the accuracy and reliability of surface water extraction. For example, Li et al. (2022) used the random forest (RF) algorithm to achieve high-precision classification of surface water in Sentinel-2 images, with producer accuracy, user accuracy and F1 value reaching 91.208%, 88.706% and 89.940% respectively. Bioresita et al. (2019) combined Sentinel-1 and Sentinel-2 images and used decision-level data fusion technology to study the extraction method of permanent and temporary water bodies in the central region of Ireland, and achieved good results. Cao et al. (2019) used the local histogram characteristics of SAR images and suppressed the influence of coherent speckle noise on water extraction results through the bimodal method. Saghafi et al. (2021) used the multiple features of Landsat-8, Sentinel-1 and Sentinel-2 images to train multiple classifiers and perform decision-level fusion to improve the accuracy of surface water identification.

[0006] Although the above studies have made significant progress in the application of remote sensing technology to surface water extraction, the following limitations still exist: (1) Many methods rely on a large number of manually drawn samples for model training, while the technology for automatically obtaining high-quality samples still needs to be further optimized; (2) The coordinated application of optical and radar data is still not in-depth enough. Most studies only use a single data source or do not fully explore the fusion potential of multi-source data; (3) Water extraction research is mostly concentrated in a single time period and lacks the support of long-term series data. As a result, the extraction results are difficult to accurately reflect the spatiotemporal dynamic changes of water bodies in cross-climatic zones, which is not conducive to in-depth analysis of the combined impact of climate and environmental changes. Summary of the invention

[0007] The technical problem to be solved by the present invention is to provide a method for active and passive collaborative remote sensing extraction of surface water resources across climate zones, which deeply integrates optical and radar data, combines automated sample generation technology and multi-time series data processing, and accurately performs time series monitoring and analysis of surface water resources.

[0008] In order to solve the above technical problems, the present invention adopts the following technical solution: the present invention designs a cross-climatic zone surface water resource active and passive collaborative remote sensing extraction method, which is used to extract the regional distribution of water bodies in the target research area, comprising the following steps:

[0009] Step A. Obtain the optical satellite historical images and radar satellite historical images of the target study area at each preset historical collection time within the preset historical period, and obtain the optical synthetic satellite image and radar synthetic satellite image corresponding to the target study area through preprocessing, and then enter step B;

[0010] Step B. Based on the optical synthetic satellite image corresponding to the target research area, and the optical synthetic satellite image corresponding to each preset first water index layer and each preset vegetation index layer, obtain the classification label of each pixel position in the satellite image under the preset water body classification and each non-water body classification, and train the target classification model to obtain the optical image water body recognition model; then enter step C;

[0011] Step C. According to the classification labels of the preset categories of each pixel position in the satellite image, the radar synthetic satellite image corresponding to the target research area and the preset second water body index value corresponding to each pixel position in the radar synthetic satellite image are used to obtain the optimal water body classification threshold, and then proceed to step D;

[0012] Step D. respectively process the local images of each pixel position in each optical satellite historical image by applying the optical image water body recognition model to obtain the classification label of each pixel position in each optical satellite historical image, and then obtain the water body distribution in each optical satellite historical image;

[0013] At the same time, for each pixel position in each radar satellite historical image, the optimal water body classification threshold is connected to compare and judge, to determine whether each pixel position in each radar satellite historical image is a water body area, and then obtain the water body distribution in each radar satellite historical image;

[0014] Then proceed to step E;

[0015] Step E. For each historical collection moment, obtain each union area between the water body distribution in the optical satellite historical image and the water body distribution in the radar satellite historical image at the historical collection moment, and each union area constitutes the water body distribution of the target study area corresponding to the historical collection moment, thereby obtaining the water body distribution of the target study area corresponding to each historical collection moment.

[0016] As a preferred technical solution of the present invention: in step A, for the obtained target study area, the optical satellite historical images and the radar satellite historical images corresponding to each preset historical collection time within the preset historical period are respectively executed, and steps A1-1 to A1-3, and steps A2-1 to A2-2 are executed simultaneously:

[0017] Step A1-1. Screen each optical satellite historical image to obtain each optical satellite historical image in which the ratio of the cloud area to the entire image area is less than a preset ratio threshold, as each optical primary satellite image, and then proceed to step A1-2;

[0018] Step A1-2. For each optical primary satellite image, remove the local cloud area covering at least a preset lower limit number of pixels in the optical primary satellite image, update the optical primary satellite image, and then proceed to step A1-3;

[0019] Step A1-3. For each pixel position in the satellite image, obtain the median of the pixel values ​​of each optical primary satellite image corresponding to the pixel position to form the pixel median corresponding to the pixel position; then combine the pixel medians corresponding to each pixel position to form an optical synthetic satellite image;

[0020] Step A2-1. For each radar satellite historical image, first perform boundary noise removal processing and update, then use the Refine Lee filtering algorithm to remove the scattered noise in the radar satellite historical image to form a radar primary satellite image, and then enter step A2-2;

[0021] Step A2-2. For each pixel position in the satellite image, obtain the median of the pixel values ​​of each radar primary satellite image corresponding to the pixel position to form the pixel median corresponding to the pixel position; then combine the pixel medians corresponding to each pixel position into a radar synthetic satellite image.

[0022] As a preferred technical solution of the present invention: the step B includes the following steps B1 to B3;

[0023] Step B1. Obtain the preset first water index values ​​corresponding to each pixel position in the optical synthetic satellite image corresponding to the target study area, including the normalized water index value NDWI, the normalized difference water index value MNDWI, the land water extraction index value LSWI, and the AWEIsh value and AWEInsh value in the automatic water extraction index, to form the first water index layers corresponding to the optical synthetic satellite image, including the normalized water index layer, the normalized difference water index layer, the land water extraction index layer, and the automatic water extraction index layer;

[0024] At the same time, the preset vegetation index values ​​corresponding to each pixel position in the optical synthetic satellite image corresponding to the target research area are obtained, including the normalized vegetation index value NDVI and the vegetation enhancement index value EVI, to form the vegetation index layers corresponding to the optical synthetic satellite image, including the normalized vegetation index layer and the vegetation enhancement index layer;

[0025] Then proceed to step B2;

[0026] Step B2. According to the optical synthetic satellite image, and the preset first water index layers and vegetation index layers corresponding to the optical synthetic satellite image, in combination with the dynamic world land classification data set, the classification labels of each pixel position in the satellite image under the preset water body classification and each non-water body classification are determined, and a preset number of pixel positions under each classification label are randomly selected to form each sample pixel position, and the sample pixel position corresponds to the local image of the optical synthetic satellite image, and the classification label of the sample pixel position is linked to form a sample, and then the sample dataset is formed by each sample, and then the process proceeds to step B3;

[0027] Step B3. Based on the sample data set, with the local image in the sample as input and the classification label in the sample as output, the target classification model is trained to obtain an optical image water body recognition model.

[0028] As a preferred technical solution of the present invention: in the step B2, for each randomly selected sample pixel position, the maximum flooding range layer in the JRC global surface water dynamic dataset is first applied to perform mask processing and update on the local image of the optical synthetic satellite image corresponding to each sample pixel position corresponding to the water body classification label, and then the local image of the optical synthetic satellite image corresponding to the sample pixel position is linked to the classification label of the sample pixel position to form a sample.

[0029] As a preferred technical solution of the present invention: the non-water body classifications include vegetation classification labels, bare land classification labels, cloud classification labels, and snow classification labels.

[0030] As a preferred technical solution of the present invention: the step C includes the following steps C1 to C3, obtaining the optimal water body classification threshold;

[0031] Step C1. Obtain a preset second water index value of a dual-polarization radar water index value SDWI corresponding to each pixel position in the radar synthetic satellite image corresponding to the target study area, and then proceed to step C2;

[0032] Step C2. According to the classification labels of each pixel position in the satellite image for each preset category, the dual-polarization radar water index value SDWI of each pixel position in the radar synthetic satellite image is counted to determine the initial water classification threshold T1 of the dual-polarization radar water index value SDWI corresponding to the pixel position in the radar synthetic satellite image to distinguish the water body classification from each non-water body classification;

[0033] At the same time, the vertical polarization backscatter coefficient value VV involved in the calculation of the dual-polarization radar water index value SDWI at each pixel position in the radar synthetic satellite image is statistically analyzed to determine the initial water body classification threshold T2 of the vertical polarization backscatter coefficient value VV corresponding to the pixel position in the radar synthetic satellite image to distinguish the water body classification from the non-water body classification;

[0034] Then proceed to step C3;

[0035] Step C3. Based on the water body initial classification threshold T1 and the water body initial classification threshold T2, respectively, according to the water body edge detection in the radar synthetic satellite image, combined with the Otsu algorithm, determine the optimal SDWI water body classification threshold for distinguishing water body classification from various non-water body classifications using the dual-polarization radar water index value SDWI corresponding to the pixel position in the radar synthetic satellite image, and the optimal VV water body classification threshold for distinguishing water body classification from various non-water body classifications using the vertical polarization backscatter coefficient value VV corresponding to the pixel position in the radar synthetic satellite image.

[0036] As a preferred technical solution of the present invention: in step C3, based on the initial classification threshold T1 of the water body, the following steps C3a-1 to C3a-3 are performed;

[0037] Step C3a-1. According to the initial classification threshold T1 of the water body, the Canny edge detection operator is applied to determine the edge information of each water body in the radar synthetic satellite image, and each water body whose edge perimeter is greater than the preset perimeter threshold is obtained as each water body to be analyzed, and then the process proceeds to step C3a-2;

[0038] Step C3a-2. For each water body to be analyzed, the edge of the water body to be analyzed is extended by a preset distance into the water and outside the water body to form an edge buffer area, thereby obtaining the edge buffer area corresponding to each water body to be analyzed, and then proceeding to step C3a-3;

[0039] Step C3a-3. Statistic the dual-polarization radar water index SDWI of each pixel position in the edge buffer area corresponding to each water body to be analyzed, and apply the Otsu algorithm to maximize the inter-class variance to determine the optimal SDWI water classification threshold for distinguishing water body classification from non-water body classifications by using the dual-polarization radar water index SDWI of the pixel position in the radar synthetic satellite image;

[0040] At the same time, based on the initial classification threshold T2 of the water body, the following steps C3b-1 to C3b-3 are performed;

[0041] Step C3b-1. According to the initial classification threshold T2 of the water body, the Canny edge detection operator is applied to determine the edge information of each water body in the radar synthetic satellite image, and each water body whose edge perimeter is greater than the preset perimeter threshold is obtained as each water body to be analyzed, and then enter step C3b-2;

[0042] Step C3b-2. For each water body to be analyzed, the edge of the water body to be analyzed is extended by a preset distance into the water and outside the water body to form an edge buffer area, thereby obtaining the edge buffer area corresponding to each water body to be analyzed, and then proceeding to step C3b-3;

[0043] Step C3b-3. The vertical polarization backscatter coefficient value VV of each pixel position in the edge buffer area corresponding to each water body to be analyzed is statistically analyzed, and the Otsu algorithm is applied to maximize the inter-class variance to determine the optimal VV water body classification threshold value for the vertical polarization backscatter coefficient value VV corresponding to the pixel position in the radar synthetic satellite image to distinguish the water body classification from the non-water body classification.

[0044] As a preferred technical solution of the present invention: in the step D, for each pixel position in each radar satellite historical image, it is determined whether the dual-polarization radar water index value SDWI of the pixel position is greater than the optimal SDWI water body classification threshold, and the vertical polarization backscatter coefficient value VV of the pixel position is less than the optimal VV water body classification threshold. If so, the pixel position is determined to be a water body area; otherwise, the pixel position is determined to be a non-water body area, so as to determine whether each pixel position in each radar satellite historical image is a water body area.

[0045] As a preferred technical solution of the present invention: it also includes step F as follows, after executing step E, entering step F;

[0046] Step F. for each union area between the water body distribution in the optical satellite historical image and the water body distribution in the radar satellite historical image at each historical collection time, perform the following operations to update the union area, and then update the water body distribution corresponding to each historical collection time in the target study area;

[0047] Firstly, the area with slope greater than the preset slope threshold is regarded as the mountain shadow area, and the union area is masked to eliminate the mountain shadow area in the union area and update the union area. Then, according to the global human settlement layer in the JRC global surface water dataset, the union area is masked again to eliminate the building shadow area and the mis-extracted building area in the union area and update the union area.

[0048] The active and passive coordinated remote sensing extraction method for surface water resources across climate zones described in the present invention has the following technical effects compared with the prior art by using the above technical solution:

[0049] (1) The present invention designs a cross-climatic zone surface water resource active and passive collaborative remote sensing extraction method. First, based on satellite historical images, optical synthetic satellite images and radar synthetic satellite images of the target study area are preprocessed, and then the optical image water body recognition model is trained and combined with the radar synthetic satellite images to obtain the optimal water body classification threshold. The water body distribution is recognized by the model recognition and the classification threshold respectively, and finally the recognition results of the two methods are integrated to obtain the water body distribution of the target study area corresponding to each historical collection time. The design scheme deeply integrates optical and radar data to achieve large-scale, long-time series rapid and automatic water system mapping, significantly improving the accuracy and reliability of the extraction results. It can not only make up for the limitations of a single data source, but also provide important support for the dynamic monitoring and scientific management of water resources in the basin. It is particularly suitable for water resources monitoring in complex basins across climate zones. In the application, it provides important data support and technical guarantee for water resources management, ecological protection and hydrological dynamic change analysis in the Nile River Basin in Africa. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 It is a schematic diagram of a cross-climatic zone active and passive coordinated remote sensing extraction method for surface water resources designed by the present invention;

[0051] Figure 2 This is an RGB composite image of the Sentinel 2 data of the Blue Nile River Basin's median composite image in 2021 obtained in an embodiment of the present invention.

[0052] Figure 3 This is a flow chart of water extraction based on Sentinel-1 remote sensing images according to an embodiment of the present invention.

[0053] Figure 4 In order to use this method to achieve the monthly water body extraction results in the Blue Nile River Basin in 2021 based on the coordination of optical and radar data. DETAILED DESCRIPTION

[0054] The specific implementation modes of the present invention will be further described in detail below in conjunction with the accompanying drawings.

[0055] The present invention designs a cross-climatic zone surface water resource active and passive collaborative remote sensing extraction method, which is used to extract the regional distribution of water bodies in the target study area. In practical applications, such as Figure 1 As shown, the specific design executes the following steps.

[0056] Step A. Obtain the optical satellite historical images and radar satellite historical images of the target study area at each preset historical collection time within the preset historical period, and obtain the optical synthetic satellite images and radar synthetic satellite images corresponding to the target study area through preprocessing, and then enter step B.

[0057] In practical applications, the above step A is to execute steps A1-1 to A1-3 for the optical satellite historical images of the target study area corresponding to each preset historical collection moment within the preset historical period, so as to obtain the optical synthetic satellite image corresponding to the target study area.

[0058] Step A1-1. Screen each optical satellite historical image to obtain each optical satellite historical image in which the cloud area accounts for a preset ratio threshold of less than 20% of the entire image area, as each optical primary satellite image, and then proceed to step A1-2.

[0059] Step A1-2. For each optical primary satellite image, remove the local cloud area covering at least a preset lower limit number of pixels in the optical primary satellite image, update the optical primary satellite image, and then proceed to step A1-3.

[0060] Step A1-3. For each pixel position in the satellite image, obtain the median of the pixel values ​​of each optical primary satellite image corresponding to the pixel position to form the pixel median corresponding to the pixel position; then combine the pixel medians corresponding to each pixel position into an optical synthetic satellite image.

[0061] At the same time, for the radar satellite historical images of the target study area corresponding to each preset historical collection moment in the preset historical period, steps A2-1 to A2-2 are executed to obtain the radar synthetic satellite images corresponding to the target study area.

[0062] Step A2-1. For each radar satellite historical image, first perform boundary noise removal processing and update, then use the Refine Lee filtering algorithm to remove the scattered noise in the radar satellite historical image to form a radar primary satellite image, and then enter step A2-2.

[0063] Step A2-2. For each pixel position in the satellite image, obtain the median of the pixel values ​​of each radar primary satellite image corresponding to the pixel position to form the pixel median corresponding to the pixel position; then combine the pixel medians corresponding to each pixel position into a radar synthetic satellite image.

[0064] Step B. Based on the optical synthetic satellite image corresponding to the target study area, and the optical synthetic satellite image corresponding to the preset first water body index layers and the preset vegetation index layers, obtain the classification label of each pixel position in the satellite image under the preset water body classification and each non-water body classification, and train the target classification model to obtain the optical image water body recognition model; then enter step C.

[0065] In actual application, the above step B is specifically designed to execute the following steps B1 to B3.

[0066] Step B1. Calculate according to the following formula:

[0067]

[0068]

[0069] AWEIsh=BLUE+2.5×GREEN-1.5×(NIR+SWIR1)-0.25×SWIR2

[0070] AWEInsh=4×(GREEN-SWIR1)-(0.25×NIR+2.75×SWIR2)

[0071] The preset first water index values ​​corresponding to each pixel position in the optical synthetic satellite image corresponding to the target study area are obtained, including the normalized water index value NDWI, the normalized difference water index value MNDWI, the land water extraction index value LSWI, and the AWEIsh value and AWEInsh value in the automatic water extraction index, to form the first water index layers corresponding to the optical synthetic satellite image, including the normalized water index layer, the normalized difference water index layer, the land water extraction index layer, and the automatic water extraction index layer. Among them, GREEN represents the pixel value of the green band corresponding to the pixel position in the optical synthetic satellite image, NIR represents the pixel value of the near infrared band corresponding to the pixel position in the optical synthetic satellite image, BLUE represents the pixel value of the blue band corresponding to the pixel position in the optical synthetic satellite image, SWIR1 and SWIR2 represent the pixel value of the shortwave infrared first band and the pixel value of the shortwave infrared second band corresponding to the pixel position in the optical synthetic satellite image, respectively.

[0072] In practical applications, the normalized difference water index value NDWI can enhance water bodies in satellite images. Its main purpose is to detect and monitor small changes in water content. The disadvantage of this index is that it is sensitive to building structures and cannot effectively suppress building noise in built-up areas and mountain shadows in mountainous areas, which may lead to an overestimation of water bodies; the normalized difference water index value MNDWI can effectively suppress background noise such as buildings and bare land, and has higher accuracy in extracting water bodies in densely built-up towns; the terrestrial water extraction index value LSWI is used to evaluate the distribution and changes of surface water bodies, which is of great significance for water resources management and environmental monitoring; the automatic water extraction index AWEI can consistently improve the extraction accuracy of water bodies in the presence of various environmental noises, while providing a stable threshold value. It has two forms, namely AWEIsh value and AWEInsh value.

[0073] At the same time, the calculation formula is as follows:

[0074]

[0075] The preset vegetation index values ​​corresponding to each pixel position in the optical synthetic satellite image corresponding to the target study area are obtained, including the normalized vegetation index value NDVI and the vegetation enhancement index value EVI, to form the vegetation index layers corresponding to the optical synthetic satellite image, including the normalized vegetation index layer and the vegetation enhancement index layer; wherein RED represents the pixel value of the red light band corresponding to the pixel position in the optical synthetic satellite image.

[0076] Then proceed to step B2.

[0077] In practical applications, the Normalized Difference Vegetation Index (NDVI) can show information such as plant growth, ecosystem vitality and productivity; the Vegetation Enhancement Index (EVI) can increase the sensitivity of vegetation, reduce soil background and atmospheric influences, and has higher sensitivity and superiority in monitoring vegetation changes. It is widely used in studies such as grassland degradation monitoring and quantitative analysis of grassland resources.

[0078] Step B2. According to the optical synthetic satellite image, the preset first water body index layers and the vegetation index layers corresponding to the optical synthetic satellite image, and in combination with the Dynamic World LandCover, the classification labels of each pixel position in the satellite image under the preset water body classification and each non-water body classification are determined, and a preset number of pixel positions under each classification label are randomly selected to form each sample pixel position; further, for each sample pixel position, the maximum flooding range layer in the JRC Global Surface Water Dynamic Dataset (JRC GSW) is first applied to perform mask processing and update on the local image of the optical synthetic satellite image corresponding to each sample pixel position corresponding to the water body classification label, and then the local image of the optical synthetic satellite image corresponding to the sample pixel position is linked to the classification label of the sample pixel position to form a sample, and then the sample dataset is formed by each sample, and then step B3 is entered. In practical applications, each non-water body classification design includes vegetation classification labels, bare land classification labels, cloud classification labels, and snow classification labels.

[0079] Step B3. Based on the sample data set, with the local image in the sample as input and the classification label in the sample as output, a target classification model such as a random forest classification model is trained to obtain an optical image water body recognition model.

[0080] Step C. According to the classification labels of the preset categories for each pixel position in the satellite image, the radar synthetic satellite image corresponding to the target research area, and the preset second water body index value corresponding to each pixel position in the radar synthetic satellite image, the optimal water body classification threshold is obtained, and then step D is entered.

[0081] In practical application, the above step C is specifically designed to execute the following steps C1 to C3 to obtain the optimal water body classification threshold.

[0082] Step C1. Calculate according to the following formula:

[0083] SDWI=ln(10×VV×VH)-8

[0084] Obtain the preset second water index value of the dual-polarization radar water index value SDWI corresponding to each pixel position in the radar synthetic satellite image corresponding to the target study area, and then enter step C2, wherein VV and VH respectively represent the vertical polarization backscatter coefficient value and the horizontal polarization backscatter coefficient value corresponding to the pixel position in the radar synthetic satellite image.

[0085] Step C2. According to the classification labels of each preset category for each pixel position in the satellite image, the dual-polarization radar water index value SDWI of each pixel position in the radar synthetic satellite image is counted to determine the initial water classification threshold T1 of the dual-polarization radar water index value SDWI corresponding to the pixel position in the radar synthetic satellite image to distinguish the water body classification from the non-water body classification.

[0086] At the same time, the vertical polarization backscatter coefficient value VV involved in the calculation of the dual-polarization radar water index value SDWI for each pixel position in the radar synthetic satellite image is statistically analyzed to determine the initial water body classification threshold T2 corresponding to the vertical polarization backscatter coefficient value VV of the pixel position in the radar synthetic satellite image to distinguish the water body classification from various non-water body classifications.

[0087] Then proceed to step C3.

[0088] The water body initial classification threshold T1 and the water body initial classification threshold T2 are obtained to ensure the high accuracy of water body sample classification, while adapting to the differences in land feature characteristics across climatic zones.

[0089] Step C3. Based on the water body initial classification threshold T1 and the water body initial classification threshold T2, respectively, according to the water body edge detection in the radar synthetic satellite image under the Canny edge detection operator, combined with the Otsu algorithm (OTSU), determine the optimal SDWI water body classification threshold for distinguishing water body classification from various non-water body classifications using the dual-polarization radar water index value SDWI corresponding to the pixel position in the radar synthetic satellite image, and the optimal VV water body classification threshold for distinguishing water body classification from various non-water body classifications using the vertical polarization backscatter coefficient value VV corresponding to the pixel position in the radar synthetic satellite image.

[0090] In practical applications, regarding the acquisition of the optimal SDWI water body classification threshold and the optimal VV water body classification threshold, the specific design is based on the initial water body classification threshold T1, and the following steps C3a-1 to C3a-3 are executed to obtain the optimal SDWI water body classification threshold.

[0091] Step C3a-1. According to the initial classification threshold T1 of the water body, the Canny edge detection operator is applied to determine the edge information of each water body in the radar synthetic satellite image, and each water body whose edge perimeter is greater than the preset perimeter threshold is obtained as each water body to be analyzed, and then enter step C3a-2.

[0092] During the execution of step C3a-1, the Canny operator obtains the transition boundary between the water body and the non-water body through gradient detection, retains key edge features, designs greater than a preset perimeter threshold, removes small area noise or invalid boundaries, and generates an optimized water body edge.

[0093] Step C3a-2. For each water body to be analyzed, the edge of the water body to be analyzed is extended by a preset distance toward the water and the outside of the water body to form an edge buffer area, thereby obtaining the edge buffer area corresponding to each water body to be analyzed, and then entering step C3a-3.

[0094] Step C3a-3. The dual-polarization radar water index value SDWI of each pixel position in the edge buffer area corresponding to each water body to be analyzed is counted, and the Otsu algorithm is applied to maximize the inter-class variance to determine the optimal SDWI water body classification threshold for distinguishing water body classification from non-water body classifications using the dual-polarization radar water index value SDWI corresponding to the pixel position in the radar synthetic satellite image.

[0095] At the same time, based on the initial water body classification threshold T2, the following steps C3b-1 to C3b-3 are performed to obtain the optimal VV water body classification threshold.

[0096] Step C3b-1. According to the initial classification threshold T2 of the water body, the Canny edge detection operator is applied to determine the edge information of each water body in the radar synthetic satellite image, and each water body whose edge perimeter is greater than the preset perimeter threshold is obtained as each water body to be analyzed, and then enter step C3b-2.

[0097] Step C3b-2. For each water body to be analyzed, the edge of the water body to be analyzed is extended by a preset distance toward the water and the outside of the water body to form an edge buffer area, thereby obtaining the edge buffer area corresponding to each water body to be analyzed, and then entering step C3b-3.

[0098] Step C3b-3. The vertical polarization backscatter coefficient value VV of each pixel position in the edge buffer area corresponding to each water body to be analyzed is statistically analyzed, and the Otsu algorithm is applied to maximize the inter-class variance to determine the optimal VV water body classification threshold value for the vertical polarization backscatter coefficient value VV corresponding to the pixel position in the radar synthetic satellite image to distinguish the water body classification from the non-water body classification.

[0099] Step D: respectively process the local images of each pixel position in each optical satellite historical image by applying the optical image water body recognition model to obtain the classification label of each pixel position in each optical satellite historical image, and then obtain the water body distribution in each optical satellite historical image.

[0100] At the same time, for each pixel position in each radar satellite historical image, a comparison is made in connection with the optimal water body classification threshold. By judging whether the dual-polarization radar water body index value SDWI of the pixel position is greater than the optimal SDWI water body classification threshold, and the vertical polarization backscatter coefficient value VV of the pixel position is less than the optimal VV water body classification threshold, if so, the pixel position is judged to be a water body area, otherwise, the pixel position is judged to be a non-water body area, and whether each pixel position in each radar satellite historical image is a water body area is determined, thereby obtaining the water body distribution in each radar satellite historical image; then entering step E.

[0101] Step E. For each historical collection moment, obtain each union area between the water body distribution in the optical satellite historical image and the water body distribution in the radar satellite historical image at the historical collection moment, and each union area constitutes the water body distribution of the target study area corresponding to the historical collection moment, and then obtain the water body distribution of the target study area corresponding to each historical collection moment, and then enter step F.

[0102] Step F. For each union area between the water body distribution in the optical satellite historical image and the water body distribution in the radar satellite historical image at each historical collection time, perform the following operations to update the union area, and then update the water body distribution corresponding to each historical collection time in the target study area.

[0103] Firstly, the terrain slope is calculated using digital elevation model (DEM) data, and the areas with slope greater than the preset slope threshold are marked as hill shadow areas. The union area is masked to remove the hill shadow areas in the union area, reduce the noise of mis-extracted water bodies, and update the union area. Then, according to the Global Human Settlements Layer (GHSL) in the JRC Global Surface Water Dataset (JRC GSW), a secondary masking process is performed on the union area to remove the building shadow areas and mis-extracted building areas in the union area, and update the union area.

[0104] The active and passive collaborative remote sensing extraction method for surface water resources across climate zones designed by the present invention is applied in practice. Specifically, based on the Google Earth Engine (GEE) cloud platform, the annual long-time series image collection of Sentinel-1GRD (radar satellite image) and Sentinel-2MSI (optical satellite image) is used to study the extraction of surface water in the Blue Nile River Basin in Africa. The advantages of optical and radar remote sensing images are fully combined, and active and passive collaborative data processing technology is used to realize the accurate extraction and dynamic monitoring of surface water in regions across climate zones.

[0105] In the selection of experimental data, Sentinel-1 (radar satellite image) and Sentinel-2 (optical satellite image) data of the Blue Nile River in Africa for the whole year of 2021 were obtained, such as Figure 2 As shown. In the Sentinel-2 image data processing, based on the cloud probability layer (S2_CLOUD_PROBABILITY) provided in the image collection, the cloud-covered areas of the Blue Nile River Basin are declouded. Through the ee.ImageCollection.median() function in GEE, the images of each month are median-synthesized to generate a monthly median image collection, namely optical synthetic satellite images and radar synthetic satellite images. At the same time, the areas with missing data and cloud-covered areas in the monthly median images are marked as Nodata areas, so that the water body information of these areas can be supplemented by Sentinel-1 images in the subsequent steps.

[0106] In the processing of Sentinel-1 (radar satellite image), the radar backscatter coefficient is first converted into dimensionless data using the formula ee.Image(10).pow(img.divide(10)) based on the GEE platform, and then the RefineLee filter is used to remove the speckle noise in the radar image; then, based on the GEE platform, the dimensionless data is converted back into backscatter coefficients using the formula ee.Image(10).multiply(img.log10()) to ensure the physical consistency of the image data. Subsequently, the ee.ImageCollection.median() function in GEE is used to generate a monthly median composite image to provide high-quality radar image data for subsequent analysis.

[0107] In order to enhance the ability to distinguish water bodies from other land features, this study selected five first water body indices (NDWI, MNDWI, LSWI, AWEIsh, AWEInsh) and two vegetation indices (NDVI, EVI). In terms of sample generation, the Dynamic World Land Cover dataset and the JRC Global Surface Water Dynamics dataset (JRC GSW) were combined as data sources for automatically generating samples. First, the Dynamic World Land Cover dataset was synthesized monthly through the ee.Reducer.mode() function in GEE; then the ee.Image.remap() function was used to reclassify the land feature classification samples in the Dynamic World Land Cover dataset into four categories: water, vegetation, land, cloud, and snow. For the reclassified dataset, 12,000 samples were randomly selected from each category as initial training samples using the ee.Image.stratifiedSample() function.

[0108] In order to further improve the quality of water samples, the water samples in the initial training samples were masked based on the maximum flooding range layer of the JRC Global Surface Water Dynamics Dataset (JRC GSW), and 10,560 high-quality water sample points were retained. After merging the water samples with other land object classification samples, the backscatter coefficient and SDWI eigenvalue of the VV polarization mode were counted. According to the distribution of eigenvalues ​​of different land object categories, the initial classification threshold of the water body in the VV polarization mode was set to -17.2, and the initial classification threshold of the SDWI water body index was set to 0.19, which were used as the initial classification criteria for the subsequent classification model.

[0109] The random forest classification method is widely used in remote sensing classification tasks due to its strong ability to resist overfitting, excellent ability to process large-scale data, and high robustness. In this study, the automatically generated sample data was input into the random forest model, and the Sentinel-2 optical image was classified to obtain the monthly water distribution results in the study area.

[0110] To supplement the water body information in the Nodata area of ​​Sentinel-2 images, Sentinel-1 data is used to extract water bodies, such as Figure 3As shown in the figure. The Edge-OTSU algorithm is used to process radar images. The algorithm automatically determines the optimal threshold based on the initial threshold to improve the accuracy of water body classification. The VV polarization data of Sentinel-1 radar images can effectively reflect the scattering characteristics of water bodies due to its strong penetration; the SDWI index combines the VV and VH polarization characteristics, further enhancing the ability to distinguish water bodies from other landforms.

[0111] Using the statistically obtained VV and SDWI initial classification values ​​as thresholds, the ee.Algorithms.CannyEdgeDetector() function in GEE is used to extract the water body edge in the Sentinel-1 image, which preliminarily represents the water body boundary in the image. Subsequently, the edge intensity threshold is set to 0.05, the water body edge information greater than the threshold is retained, the noise and weak boundary information are removed, and the optimized Sentinel-1 water body boundary is generated. The optimized boundary is buffered, the buffer width is set to 500 meters, and the histogram distribution is statistically analyzed within the buffer. Based on the OTSU algorithm, the VV polarization and SDWI water body index are respectively classified for water bodies, and the intersection of the two extraction results is taken to finally obtain the water body extraction result of the Sentinel-1 image.

[0112] In the fusion stage of water extraction results, the water extraction results of Sentinel-1 images are merged with those of Sentinel-2 images to generate seamless surface water distribution data of the Blue Nile River Basin. To further improve the accuracy of the results, the slope threshold is set to 5 degrees, and the slope is calculated by the ee.Terrain.slope() function in GEE using the DEM data of the Shuttle Radar Topography Mission (SRTM). Areas with slopes greater than 5 degrees are masked as mountain shadow areas to remove possible erroneous water results. In addition, the water extraction results are masked twice in combination with the Global Human Settlements Layer (GHSL) to remove erroneous areas caused by building shadows.

[0113] Finally, high-precision monthly water extraction results were obtained using the coordinated Sentinel-1 and Sentinel-2 data, such as Figure 4 As shown, the results have achieved seamless and dynamic water distribution monitoring in a complex environment across climate zones, providing solid data support and technical guarantee for water resources management and dynamic change analysis in the Blue Nile River Basin in Africa.

[0114] The active and passive collaborative remote sensing extraction method for surface water resources across climate zones designed by the present invention firstly takes historical satellite images as a basis, pre-processes optical synthetic satellite images and radar synthetic satellite images of the target study area, then trains to obtain an optical image water body recognition model, and links the radar synthetic satellite images to obtain the optimal water body classification threshold, and respectively realizes the recognition of water body distribution by model recognition and classification threshold, and finally integrates the recognition results of the two methods to obtain the water body distribution of the target study area corresponding to each historical collection time; the design scheme deeply integrates optical and radar data, realizes large-scale, long-time series rapid and automatic water system mapping, significantly improves the accuracy and reliability of the extraction results, can not only make up for the limitations of a single data source, but also can provide important support for the dynamic monitoring and scientific management of water resources in the basin, and is particularly suitable for water resources monitoring in complex basins across climate zones. In applications, it provides important data support and technical guarantee for water resources management, ecological protection and hydrological dynamic change analysis in the Nile River Basin in Africa.

[0115] The embodiments of the present invention are described in detail above with reference to the accompanying drawings, but the present invention is not limited to the above embodiments, and various changes can be made within the knowledge scope of ordinary technicians in this field without departing from the purpose of the present invention.

Claims

1. A cross-climatic zone active and passive collaborative remote sensing extraction method for surface water resources, used to extract the regional distribution of water bodies in the target study area, characterized in that: The steps include: Step A. Obtain the optical satellite historical images and radar satellite historical images of the target study area at each preset historical collection time within the preset historical period, and obtain the optical synthetic satellite image and radar synthetic satellite image corresponding to the target study area through preprocessing, and then enter step B; Step B. Based on the optical synthetic satellite image corresponding to the target research area, and the optical synthetic satellite image corresponding to each preset first water index layer and each preset vegetation index layer, obtain the classification label of each pixel position in the satellite image under the preset water body classification and each non-water body classification, and train the target classification model to obtain the optical image water body recognition model; then enter step C; Step C. According to the classification labels of the preset categories of each pixel position in the satellite image, the radar synthetic satellite image corresponding to the target research area and the preset second water body index value corresponding to each pixel position in the radar synthetic satellite image are used to obtain the optimal water body classification threshold, and then proceed to step D; Step D. respectively process the local images of each pixel position in each optical satellite historical image by applying the optical image water body recognition model to obtain the classification label of each pixel position in each optical satellite historical image, and then obtain the water body distribution in each optical satellite historical image; At the same time, for each pixel position in each radar satellite historical image, the optimal water body classification threshold is connected to compare and judge, to determine whether each pixel position in each radar satellite historical image is a water body area, and then obtain the water body distribution in each radar satellite historical image; Then proceed to step E; Step E. For each historical collection moment, obtain each union area between the water body distribution in the optical satellite historical image and the water body distribution in the radar satellite historical image at the historical collection moment, and each union area constitutes the water body distribution of the target study area corresponding to the historical collection moment, thereby obtaining the water body distribution of the target study area corresponding to each historical collection moment.

2. According to claim 1, a cross-climate zone active and passive collaborative remote sensing extraction method for surface water resources is characterized by: In step A, for the obtained target research area, the optical satellite historical images and the radar satellite historical images corresponding to each preset historical collection time within the preset historical period are respectively executed, and steps A1-1 to A1-3, and steps A2-1 to A2-2 are simultaneously executed: Step A1-1. Screen each optical satellite historical image to obtain each optical satellite historical image in which the ratio of the cloud area to the entire image area is less than a preset ratio threshold, as each optical primary satellite image, and then proceed to step A1-2; Step A1-2. For each optical primary satellite image, remove the local cloud area covering at least a preset lower limit number of pixels in the optical primary satellite image, update the optical primary satellite image, and then proceed to step A1-3; Step A1-3. For each pixel position in the satellite image, obtain the median of the pixel values ​​of each optical primary satellite image corresponding to the pixel position to form the pixel median corresponding to the pixel position; then combine the pixel medians corresponding to each pixel position to form an optical synthetic satellite image; Step A2-1. For each radar satellite historical image, first perform boundary noise removal processing and update, then use the Refine Lee filtering algorithm to remove the scattered noise in the radar satellite historical image to form a radar primary satellite image, and then enter step A2-2; Step A2-2. For each pixel position in the satellite image, obtain the median of the pixel values ​​of each radar primary satellite image corresponding to the pixel position to form the pixel median corresponding to the pixel position; then combine the pixel medians corresponding to each pixel position into a radar synthetic satellite image.

3. According to claim 1, a cross-climate zone active and passive collaborative remote sensing extraction method for surface water resources is characterized by: The step B includes the following steps B1 to B3; Step B1. Obtain the preset first water index values ​​corresponding to each pixel position in the optical synthetic satellite image corresponding to the target study area, including the normalized water index value NDWI, the normalized difference water index value MNDWI, the land water extraction index value LSWI, and the AWEIsh value and AWEInsh value in the automatic water extraction index, to form the first water index layers corresponding to the optical synthetic satellite image, including the normalized water index layer, the normalized difference water index layer, the land water extraction index layer, and the automatic water extraction index layer; At the same time, the preset vegetation index values ​​corresponding to each pixel position in the optical synthetic satellite image corresponding to the target research area are obtained, including the normalized vegetation index value NDVI and the vegetation enhancement index value EVI, to form the vegetation index layers corresponding to the optical synthetic satellite image, including the normalized vegetation index layer and the vegetation enhancement index layer; Then proceed to step B2; Step B2. According to the optical synthetic satellite image, and the preset first water index layers and vegetation index layers corresponding to the optical synthetic satellite image, in combination with the dynamic world land classification data set, the classification labels of each pixel position in the satellite image under the preset water body classification and each non-water body classification are determined, and a preset number of pixel positions under each classification label are randomly selected to form each sample pixel position, and the sample pixel position corresponds to the local image of the optical synthetic satellite image, and the classification label of the sample pixel position is linked to form a sample, and then the sample dataset is formed by each sample, and then the process proceeds to step B3; Step B3. Based on the sample data set, with the local image in the sample as input and the classification label in the sample as output, the target classification model is trained to obtain an optical image water body recognition model.

4. According to claim 3, a cross-climatic zone surface water resource active and passive coordinated remote sensing extraction method is characterized by: In step B2, for each randomly selected sample pixel position, the maximum flooding range layer in the JRC global surface water dynamics dataset is first applied to perform mask processing and update on the local image of the optical synthetic satellite image corresponding to each sample pixel position corresponding to the water body classification label, and then the local image of the optical synthetic satellite image corresponding to the sample pixel position is linked to the classification label of the sample pixel position to form a sample.

5. The method for extracting surface water resources across climate zones by active and passive coordinated remote sensing according to claim 3, characterized in that: The non-water body classifications include vegetation classification labels, bare land classification labels, cloud classification labels, and snow classification labels.

6. The method for extracting surface water resources across climate zones by active and passive coordinated remote sensing according to claim 1, characterized in that: The step C includes the following steps C1 to C3, obtaining the optimal water body classification threshold; Step C1. Obtain a preset second water index value of a dual-polarization radar water index value SDWI corresponding to each pixel position in the radar synthetic satellite image corresponding to the target study area, and then proceed to step C2; Step C2. According to the classification labels of each pixel position in the satellite image for each preset category, the dual-polarization radar water index value SDWI of each pixel position in the radar synthetic satellite image is counted to determine the initial water classification threshold T1 of the dual-polarization radar water index value SDWI corresponding to the pixel position in the radar synthetic satellite image to distinguish the water body classification from each non-water body classification; At the same time, the vertical polarization backscatter coefficient value VV involved in the calculation of the dual-polarization radar water index value SDWI at each pixel position in the radar synthetic satellite image is statistically analyzed to determine the initial water body classification threshold T2 of the vertical polarization backscatter coefficient value VV corresponding to the pixel position in the radar synthetic satellite image to distinguish the water body classification from the non-water body classification; Then proceed to step C3; Step C3. Based on the water body initial classification threshold T1 and the water body initial classification threshold T2, respectively, according to the water body edge detection in the radar synthetic satellite image, combined with the Otsu algorithm, determine the optimal SDWI water body classification threshold for distinguishing water body classification from various non-water body classifications using the dual-polarization radar water index value SDWI corresponding to the pixel position in the radar synthetic satellite image, and the optimal VV water body classification threshold for distinguishing water body classification from various non-water body classifications using the vertical polarization backscatter coefficient value VV corresponding to the pixel position in the radar synthetic satellite image.

7. The method for extracting surface water resources across climate zones by active and passive coordinated remote sensing according to claim 6, characterized in that: In the step C3, based on the initial classification threshold T1 of the water body, the following steps C3a-1 to C3a-3 are performed; Step C3a-1. According to the initial classification threshold T1 of the water body, the Canny edge detection operator is applied to determine the edge information of each water body in the radar synthetic satellite image, and each water body whose edge perimeter is greater than the preset perimeter threshold is obtained as each water body to be analyzed, and then the process proceeds to step C3a-2; Step C3a-2. For each water body to be analyzed, the edge of the water body to be analyzed is extended by a preset distance into the water and outside the water body to form an edge buffer area, thereby obtaining the edge buffer area corresponding to each water body to be analyzed, and then proceeding to step C3a-3; Step C3a-3. Statistic the dual-polarization radar water index SDWI of each pixel position in the edge buffer area corresponding to each water body to be analyzed, and apply the Otsu algorithm to maximize the inter-class variance to determine the optimal SDWI water classification threshold for distinguishing water body classification from non-water body classifications by using the dual-polarization radar water index SDWI of the pixel position in the radar synthetic satellite image; At the same time, based on the initial classification threshold T2 of the water body, the following steps C3b-1 to C3b-3 are performed; Step C3b-1. According to the initial classification threshold T2 of the water body, the Canny edge detection operator is applied to determine the edge information of each water body in the radar synthetic satellite image, and each water body whose edge perimeter is greater than the preset perimeter threshold is obtained as each water body to be analyzed, and then enter step C3b-2; Step C3b-2. For each water body to be analyzed, the edge of the water body to be analyzed is extended by a preset distance into the water and outside the water body to form an edge buffer area, thereby obtaining the edge buffer area corresponding to each water body to be analyzed, and then proceeding to step C3b-3; Step C3b-3. The vertical polarization backscatter coefficient value VV of each pixel position in the edge buffer area corresponding to each water body to be analyzed is statistically analyzed, and the Otsu algorithm is applied to maximize the inter-class variance to determine the optimal VV water body classification threshold value for the vertical polarization backscatter coefficient value VV corresponding to the pixel position in the radar synthetic satellite image to distinguish the water body classification from the non-water body classification.

8. The method for extracting surface water resources across climate zones by active and passive coordinated remote sensing according to claim 6 or 7, characterized in that: In the step D, for each pixel position in each radar satellite historical image, it is determined whether the dual-polarization radar water index value SDWI of the pixel position is greater than the optimal SDWI water body classification threshold, and the vertical polarization backscatter coefficient value VV of the pixel position is less than the optimal VV water body classification threshold. If so, the pixel position is determined to be a water body area; otherwise, the pixel position is determined to be a non-water body area, so as to determine whether each pixel position in each radar satellite historical image is a water body area.

9. The method for extracting surface water resources across climate zones by active and passive coordinated remote sensing according to claim 1, characterized in that: The step F is as follows: after executing step E, entering step F; Step F. for each union area between the water body distribution in the optical satellite historical image and the water body distribution in the radar satellite historical image at each historical collection time, perform the following operations to update the union area, and then update the water body distribution corresponding to each historical collection time in the target study area; Firstly, the area with slope greater than the preset slope threshold is regarded as the mountain shadow area, and the union area is masked to eliminate the mountain shadow area in the union area and update the union area. Then, according to the global human settlement layer in the JRC global surface water dataset, the union area is masked again to eliminate the building shadow area and the mis-extracted building area in the union area and update the union area.

Images