A method for constructing a sample library of intelligent recognition of surface water bodies based on remote sensing images

Abstract

The application provides a kind of based on remote sensing image ground surface water body intelligent identification sample library construction method, comprising obtaining land use data and satellite remote sensing image;From the land use data, the polygon of ground surface water body and the corresponding water body center point are extracted, each water body center point represents a water body;According to the distribution of the water body center point, different size quadrat area is drawn;Select the water body in quadrat area as sample;The picture of each sample is generated using the satellite remote sensing image;According to the remote sensing image classification feature and polygon shape classification feature, sample classification is carried out;The information table of all samples is established.The application proposes a kind of based on remote sensing image ground surface water body intelligent identification sample library construction method, based on the highest resolution land use data and online high resolution remote sensing data in the world, combined with the distribution characteristics of ground surface water body, the intelligent identification of ground surface water body is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of remote sensing target recognition, and in particular to a method for constructing a ground water body intelligent recognition sample library based on remote sensing images. BACKGROUND

[0002] In recent years, with the development of earth observation technology, remote sensing image data has grown at a geometric rate, and higher requirements for fast and accurate interpretation methods and technologies have been put forward. The contradiction between the large accumulation of original shooting data and the insufficient extraction of available information is becoming increasingly prominent. A large amount of multi-type remote sensing image sample library is the basis for realizing high-precision intelligent interpretation of large-scale remote sensing images. At present, remote sensing interpretation sample data is constructed for different interpretation tasks, mainly including scene classification samples, ground object detection samples, ground feature classification samples, and change detection samples. In recent years, deep learning technology has been applied to scene understanding, ground object detection, and land cover classification tasks, which has promoted the rapid development of remote sensing classification sample data acquisition, labeling, and application. By constructing a large number of sample data to train deep learning networks, the feature extraction effect of remote sensing images has been significantly improved. However, overall, the practicality and commercialization of intelligent remote sensing interpretation systems have not yet reached the level of ordinary image interpretation such as face and fingerprint recognition.

[0003] Ground water bodies, as a very typical ground feature, are influenced by factors such as scale, terrain, rainfall, and economic development level, resulting in heterogeneity of the same type of water objects in different regions in terms of remote sensing image features. This results in high labeling costs, non-uniform classification systems, and a small number of existing samples, making it difficult to support accurate identification and classification of a large number of water objects in large-scale images.

[0004] The invention patent application with the application number CN 119723332 A discloses a remote sensing image ground water body extraction method and device, system, and storage medium, which includes the following steps: step S1, acquiring remote sensing ground water body data set; step S2, constructing remote sensing ground water body extraction network RSWE-Net according to the remote sensing ground water body data set; step S3, dividing the remote sensing ground water body data set to obtain a training set and a test set; step S4, training the remote sensing ground water body extraction network RSWE-Net according to the training set; and step S5, inputting the test set into the trained remote sensing ground water body extraction network RSWE-Net to extract the ground water body in the remote sensing image. The disadvantage of this method is that it only extracts the water body from the remote sensing image, and cannot classify the ground water body.

[0005] The invention patent application with the application number CN 118736438 A discloses a surface water body extraction method based on satellite hyperspectral remote sensing data, which comprises the following main steps: step 1, data preprocessing of the obtained multiple satellite hyperspectral remote sensing data; step 2, selecting feature pixels of different ground objects on the image, and constructing ground object spectral feature curves; step 3, analyzing the spectral feature curves of different types of ground objects, constructing a new water body extraction model, and processing the image using the new model; step 4, constructing an algorithm to separate water bodies and other ground objects from the above processed image. The disadvantage of this method is that it can extract water bodies from hyperspectral remote sensing data and distinguish water bodies from other ground objects, but it cannot distinguish which type of water body it is. SUMMARY

[0006] To solve the above technical problems, the present application provides a kind of based on remote sensing image ground surface water body intelligent identification sample library construction method, based on the highest resolution land use data and online high-resolution remote sensing data of global public, in combination with the distribution characteristics of ground surface water body, realize the intelligent identification of ground surface water body.

[0007] The purpose of the present application is to provide a kind of based on remote sensing image ground surface water body intelligent identification sample library construction method, including obtaining land use data and satellite remote sensing image, also includes the following steps:

[0008] Step 1: extracting ground surface water body polygons and corresponding water body center points from the land use data, each water body center point representing a water body;

[0009] Step 2: drawing different size quadrats according to the distribution of the water body center points;

[0010] Step 3: selecting water bodies within the quadrat as samples;

[0011] Step 4: generating a picture of each sample using the satellite remote sensing image;

[0012] Step 5: classifying samples according to remote sensing image classification features and polygon shape classification features;

[0013] Step 6: establishing an information table of all samples.

[0014] Preferably, step 1 includes extracting water body grids from 10m resolution land use raster data, converting the obtained water body grids to vector polygons using ArcGIS software, and saving the polygons as shp format polygon files; using the FeatureToPoint tool to extract the center point coordinates of each polygon, saving the shp format point file, which represents the geographical location of the corresponding water body polygon.

[0015] In any of the above schemes, preferably, the step 2 comprises setting sample plot center points at intervals of 5° longitude x 5° latitude, and searching for water body center points around each sample plot center point from near to far, and taking the longitude and latitude distance when the number of water bodies reaches 100 as the sample plot area radius, if the sample plot area radius is 5 (unit: degree) and the number of water bodies is less than 100, then taking 5 as the sample plot area radius, and the formula is

[0016]

[0017] wherein, R i is the radius of the sample plot area, i is the number of water bodies in the sample plot area when the radius is 5, n 100 is the distance from the sample plot center point of the sample plot area to the farthest water body point from the sample plot center point among the 100 water bodies closest to the sample plot center point. l i i,j is the distance from the water body to the sample plot area center point,

[0018]

[0019]

[0020] wherein, l i and j i are the longitude and latitude coordinates of the sample plot center point, respectively, i j and x j are the longitude and latitude coordinates of the water body center point, respectively. y x y

[0021] In any of the above schemes, preferably, the step 2 further comprises drawing a buffer area for each sample plot area with the center point of each sample plot area as the center and the sample plot area radius as the distance, to obtain circular sample plot areas for all samples.

[0022] In any of the above schemes, preferably, the step 3 comprises using the circular sample plot areas and the water body center point layer to perform the Identity tool in ArcGIS software, extracting the sample plot numbers into the water body center point layer, and using the join and Calculate Field tools to associate the sample plot numbers to the water body polygon layer, so that the water body polygon with the sample plot number is the selected water body sample.

[0023] In any of the above schemes, preferably, the step 4 comprises the following sub-steps:​​​​​

[0024] Step 41: Load the sample point layer into ArcGIS software, load the online high-resolution imagery World Imagery published by ESRI as the background, set the water sample polygons to be hollow, and only display the boundaries;

[0025] Step 42: Write a batch processing program to generate a PNG image for each water sample using ArcGIS software.

[0026] In any of the above schemes, the generation method preferably includes the following sub-steps:

[0027] Step 421: Obtain each water body polygon object;

[0028] Step 422: Obtain the bounding box of the water body polygon;

[0029] Step 423: Enlarge the bounding box of the water body polygon;

[0030] Step 424: Assign the coordinates of the enlarged water body polygon bounding box to the map;

[0031] Step 425: Set the image export path;

[0032] Step 426: Execute the image export command.

[0033] In any of the above schemes, it is preferred that the types of sample classification include: reservoirs, ponds, rivers, lakes, artificial ponds, natural ponds, ditches, and temporary water bodies.

[0034] In any of the above schemes, the preferred option is that the remote sensing image classification features of each type of sample are:

[0035] 1) Reservoir: There are clearly visible water-blocking facilities on one side of the water body, and it has obvious characteristics of intercepting the river;

[0036] 2) Ponds and dams: These have obvious characteristics of intercepting rivers, but the dam body is not obvious, and they are generally small in area;

[0037] 3) River: A section of a river clearly visible on the remote sensing image;

[0038] 4) Lakes: There are no artificial facilities around the sample, and the shoreline is naturally formed;

[0039] 5) Artificial pits / ponds: The area around the sample consists of artificially constructed water-retaining structures;

[0040] 6) Natural ponds: Independent bodies of water surrounded by artificial structures; the sample area is irregular.

[0041] 7) Ditches: Man-made channels can be seen on both sides of the water body.

[0042] 8) Temporary water bodies: river beaches and low-lying areas with standing water.

[0043] In any of the above schemes, the preferred option is that the polygon shape classification feature of each type of sample is:

[0044] 1) Reservoirs: The sample shape is generally an asymmetrical polygon;

[0045] 2) Ponds and dams: shaped like reservoirs, but small in area;

[0046] 3) Rivers: The samples were mainly long and slender in shape;

[0047] 4) Lakes: mainly circular or oval in shape, with smooth edges, located at the lowest point of the terrain;

[0048] 5) Artificial pits: square, rectangular, circular, or elliptical, with smooth edges;

[0049] 6) Natural pits and ponds: with rough boundaries and irregular shapes;

[0050] 7) Ditch: The shape is straight and the width is uniform;

[0051] 8) Temporary water bodies: irregular in shape.

[0052] In any of the above schemes, step 6 preferably includes standardizing the sample results to form a water body classification remote sensing sample library.

[0053] In any of the above solutions, step 6 preferably includes the following sub-steps:

[0054] Step 61: Based on the classification results, establish a sample information table, including sample number, sample type, longitude of the center point, latitude of the center point, and water area;

[0055] Step 62: Store the samples in folders according to their type and number them sequentially. Reservoirs, ponds, rivers, lakes, artificial ponds, natural ponds, ditches, and temporary water bodies are numbered a, b, c, d, e, f, g, and h respectively, and then numbered sequentially using 6-digit numbers.

[0056] This invention proposes a method for constructing a sample library for intelligent identification of surface water bodies based on remote sensing imagery. Based on globally available high-resolution land use data and online high-resolution remote sensing data, and combined with the distribution characteristics of surface water bodies, a sample library for intelligent identification of surface water bodies is constructed, which can provide samples for training intelligent models for refined classification of surface water bodies. Attached Figure Description

[0057] Figure 1 This is a flowchart of a preferred embodiment of the method for constructing a sample library for intelligent identification of surface water bodies based on remote sensing images according to the present invention.

[0058] Figure 2 This is a schematic diagram of an embodiment of the sample point images of the intelligent identification sample library construction method for surface water bodies based on remote sensing images according to the present invention.

[0059] Figure 3 This is a schematic diagram of a reservoir according to a preferred embodiment of the method for constructing a sample library for intelligent identification of surface water bodies based on remote sensing images according to the present invention.

[0060] Figure 4 This is a schematic diagram of a pond dam according to a preferred embodiment of the method for constructing a sample library for intelligent identification of surface water bodies based on remote sensing images according to the present invention.

[0061] Figure 5 This is a schematic diagram of a river according to a preferred embodiment of the method for constructing a sample library for intelligent identification of surface water bodies based on remote sensing images according to the present invention.

[0062] Figure 6 This is a schematic diagram of a lake according to a preferred embodiment of the method for constructing a sample library for intelligent identification of surface water bodies based on remote sensing images according to the present invention.

[0063] Figure 7 This is a schematic diagram of an artificial pond according to a preferred embodiment of the method for constructing a sample library for intelligent identification of surface water bodies based on remote sensing images according to the present invention.

[0064] Figure 8 This is a schematic diagram of a natural pond according to a preferred embodiment of the method for constructing a sample library for intelligent identification of surface water bodies based on remote sensing images according to the present invention.

[0065] Figure 9 This is a schematic diagram of a ditch according to a preferred embodiment of the method for constructing a sample library for intelligent identification of surface water bodies based on remote sensing images according to the present invention.

[0066] Figure 10 This is a temporary water body schematic diagram of a preferred embodiment of the method for constructing a sample library for intelligent identification of surface water bodies based on remote sensing images according to the present invention. Detailed Implementation

[0067] The present invention will be further described below with reference to the accompanying drawings and specific embodiments.

[0068] Example 1

[0069] like Figure 1 As shown, a method for constructing a sample library for intelligent identification of surface water bodies based on remote sensing images is described. Step 1000 involves acquiring land use data and satellite remote sensing images.

[0070] Execute step 1100: extract surface water body polygons and corresponding water body center points from the land use data. Each water body center point represents a water body. Extract rasters with the attribute of water body from the 10m resolution land use raster data. Convert the obtained water body rasters into vector polygons using ArcGIS software and save them as shapefile polygon files. Use the FeatureToPoint tool to extract the coordinates of the center point of each polygon and save them as shapefile point files to represent the geographical location of the corresponding water body polygon.

[0071] Execute step 1200: Based on the distribution of the water body center points, draw quadrat areas of different sizes. Set quadrat center points at intervals of 5° longitude × 5° latitude, and search for the water body center points surrounding each quadrat center point from near to far. Use the longitude and latitude distance when the number of water bodies reaches 100 as the quadrat area radius. If the quadrat area radius is 5 (in degrees) and the number of water bodies is less than 100, then use 5 as the quadrat area radius. The formula is as follows:

[0072]

[0073] in, R i For sample plots i radius, n This represents the number of water bodies within a quadrat with a radius of 5. l 100 For the sample plot area i The distance from the center point of the quadrat to the farthest water body among the 100 water bodies closest to the center point of the quadrat.

[0074] The formula for calculating the distance from the water body to the center point of the quadrat is:

[0075]

[0076] in, l i,j water body j to the sample plot area i Distance from the center point x i and y i These are the latitude and longitude coordinates of the center point of the quadrat. x j and y j These are the latitude and longitude coordinates of the center point of the water body.

[0077] A buffer zone for each quadrat is drawn with the center point of each quadrat as the center and the radius of the quadrat as the distance, resulting in circular quadrat zones for all quadrats.

[0078] In step 1300, select the water body within the quadrat area as the sample. Use the Identity tool in ArcGIS software to extract the quadrat number into the water body center point layer using the circular quadrat area and water body center point layer. Use the join and Calculate Field tools to associate the quadrat number with the water body polygon layer. The water body polygon with the obtained quadrat number is the selected water body sample.

[0079] Step 1400 involves generating an image for each sample using the satellite remote sensing imagery, including the following sub-steps:

[0080] Execute step 1410 to load the sample point layer into ArcGIS software, load the online high-resolution imagery World Imagery published by ESRI as the background, set the water sample polygons to be hollow, and only display the boundaries;

[0081] Execute step 1420 to write a batch program and use ArcGIS software to generate a PNG image for each water sample.

[0082] The generation method includes the following sub-steps:

[0083] Perform step 1421 to obtain each water body polygon object;

[0084] Perform step 1422 to obtain the bounding box of the water body polygon;

[0085] Perform step 1423 to enlarge the bounding box of the water body polygon;

[0086] Execute step 1424 to assign the coordinates of the enlarged water body polygon bounding box to the map;

[0087] Perform step 1425 to set the image export path;

[0088] Execute step 1426 to execute the image export command.

[0089] Step 1500 involves classifying samples based on remote sensing image classification features and polygon shape classification features. The types of sample classification include: reservoirs, ponds, rivers, lakes, artificial ponds, natural ponds, ditches, and temporary water bodies.

[0090] The remote sensing image classification features of various types of samples are as follows:

[0091] 1) Reservoir: There are clearly visible water-blocking facilities on one side of the water body, and it has obvious characteristics of intercepting the river;

[0092] 2) Ponds and dams: These have obvious characteristics of intercepting rivers, but the dam body is not obvious, and they are generally small in area;

[0093] 3) River: A section of a river clearly visible on the remote sensing image;

[0094] 4) Lakes: There are no artificial facilities around the sample, and the shoreline is naturally formed;

[0095] 5) Artificial pits / ponds: The area around the sample consists of artificially constructed water-retaining structures;

[0096] 6) Natural ponds: Independent bodies of water surrounded by artificial structures; the sample area is irregular.

[0097] 7) Ditches: Man-made channels can be seen on both sides of the water body.

[0098] 8) Temporary water bodies: river beaches and low-lying areas with standing water.

[0099] The polygon shape classification features of various types of samples are as follows

[0100] 1) Reservoirs: The sample shape is generally an asymmetrical polygon;

[0101] 2) Ponds and dams: shaped like reservoirs, but small in area;

[0102] 3) Rivers: The samples were mainly long and slender in shape;

[0103] 4) Lakes: mainly circular or oval in shape, with smooth edges, located at the lowest point of the terrain;

[0104] 5) Artificial pits: square, rectangular, circular, or elliptical, with smooth edges;

[0105] 6) Natural pits and ponds: with rough boundaries and irregular shapes;

[0106] 7) Ditch: The shape is straight and the width is uniform;

[0107] 8) Temporary water bodies: irregular in shape.

[0108] Execute step 1600 to create an information table for all samples, including the following sub-steps:

[0109] Execute step 1610 to establish a sample information table based on the classification results, including sample number, sample type, longitude of center point, latitude of center point, and water area;

[0110] Perform step 1620, store the samples in folders according to type and number them sequentially. Reservoirs, ponds, rivers, lakes, artificial ponds, natural ponds, ditches, and temporary water bodies are numbered a, b, c, d, e, f, g, and h respectively, and then numbered sequentially using 6-digit numbers.

[0111] Example 2

[0112] This paper proposes a method for constructing a sample database based on high-resolution land use data and remote sensing imagery. The process is as follows:

[0113] Step 1: Obtain the polygons of surface water bodies and their corresponding center points, with each center point representing a water body. Specifically, extract the raster with the attribute of water body from the 10m resolution land use raster data. Convert the obtained water body raster into vector polygons using ArcGIS software and save them as shapefile (shp) format polygon files. Use the FeatureToPoint tool to extract the coordinates of the center point of each polygon and save them as shapefile (shp) format point files, which will be used to represent the geographical location of the corresponding water body polygon.

[0114] Step 2: Globally, draw quadrat plots of different sizes based on the distribution of water body center points. Considering the differences in topography, rainfall, and water resource utilization across different regions globally, the characteristics of similar water bodies in imagery are inconsistent. Therefore, quadrat plots of different scales are used in different regions based on the sparse distribution of samples. Specifically, quadrat center points are set at 5° (longitude) × 5° (latitude) intervals, and the center points of water bodies surrounding each quadrat center point are searched from near to far. The latitude and longitude distance when the number of water bodies reaches 100 is used as the quadrat plot radius. If the number of water bodies is less than 100 when the quadrat plot radius is 5 (in degrees), then 5 is used as the quadrat plot radius. The formula is as follows:

[0115]

[0116] In the formula, R i Indicates the sample plot area i radius, n This indicates the amount of water within a quadrat with a radius of 5. l 100 Indicates distance from the quadrat area i The distance from the center point of the quadrat to the farthest water body among the 100 water bodies closest to the center point of the quadrat.

[0117] The method for calculating the distance from the water body to the center point of the sample plot is as follows:

[0118]

[0119] In the formula, l i,j Indicates water body j to the sample plot area i Distance from the center point x i , y i These represent the latitude and longitude coordinates of the center point of the quadrat, respectively. x j , yj These represent the latitude and longitude coordinates of the center point of the water body.

[0120] A buffer zone for each quadrat is drawn with the center point of each quadrat as the center and the radius of the quadrat as the distance, resulting in a circular area for all quadrats.

[0121] Step 3: Determine the sample points. Using the circular quadrat area drawn in the previous step and the water body center point layer obtained in the first step, use the Identity tool in ArcGIS to extract the quadrat numbers into the water body center point layer. Then, use the Join and Calculate Field tools to associate the quadrat numbers with the water body polygon layer. The water body polygon with the quadrat number is the selected water body sample.

[0122] Step 4: Extract the remote sensing image for each sample. Specifically, load the sample point layer into ArcGIS software, use the publicly available online high-resolution imagery World Imagery from ESRI as the background, and set the water sample polygons to be hollow, displaying only the boundaries. Write a batch processing program to generate a PNG image for each water sample using ArcGIS software. The generation method is as follows: first, obtain the bounding rectangle of the sample polygon, and then enlarge the rectangle by 1.5 times. This ensures that the sample polygon and surrounding image are clearly displayed in the image, facilitating the subsequent classification and labeling.

[0123] The core code for exporting images is as follows:

[0124] Get each water body polygon object: Set pF = pFClass.GetFeature(i);

[0125] Get the bounding box of the water body polygon: Set pEnvelope = pF.Shape.Envelope;

[0126] Enlarge the bounding box of the water body polygon: pEnvelope.Expand 1.5, 1.5, True;

[0127] Assign the coordinates of the enlarged water body polygon bounding box to the map: pPixelBoundsEnv.PutCoordsexportRECT.Left, exportRECT.Top, exportRECT.Right, exportRECT.bottom;

[0128] pExport.PixelBounds = pPixelBoundsEnv;

[0129] Set the image export path: pExport.ExportFileName = "D:\pic1\" + picName + ".png";

[0130] Execute the image export command: pActiveView.Output hDC, 100, exportRECT, Nothing,Nothing.

[0131] Step 5 involves sample classification. Water bodies are generally classified according to industry standards into rivers, lakes, reservoirs, ponds, ditches, and glaciers. However, based on remote sensing image features, clear distinctions are difficult. For example, my country defines a reservoir as "the water surface enclosed by the shoreline at the normal storage level of a reservoir with a total designed storage capacity ≥ 100,000 cubic meters, formed by artificial interception and collection." However, the storage capacity cannot be clearly distinguished in remote sensing images. Therefore, this application combines the features of water body remote sensing images and the polygonal characteristics of water bodies to reclassify them, dividing the samples into eight categories: reservoirs, ponds, rivers, lakes, artificial ponds, natural ponds, ditches, and temporary water bodies. The image classification features and shape features of each category are redefined. Table 1 shows the definition of each sample type and the features on the image samples. During the classification process, a small number of blurry or poor-quality samples obscured by clouds or fog are removed.

[0132] Table 1 Sample definition and classification features

[0133]

[0134] Step 6 involves standardizing the sample results to create a water body classification remote sensing sample library. Based on the classification results, a sample information table is established, including sample number, sample type, center point longitude, center point latitude, and water body area. Samples are stored in folders according to type and numbered sequentially. Reservoirs, ponds, rivers, lakes, artificial ponds, natural ponds, ditches, and temporary water bodies are numbered a, b, c, d, e, f, g, and h respectively, and then sequentially numbered using 6-digit numbers. For example, the first reservoir sample is numbered a000001.

[0135] Example 3

[0136] (1) Extract the location of the center point of the water body from the global 2023 10-meter resolution land use data.

[0137] (2) Draw the range of quadrat areas with different radii based on the distribution of the center point of the water body.

[0138] (3) Select the water body in each quadrat area as the sample point.

[0139] (4) such as Figure 2As shown, the sample point layer is loaded into ArcGIS software, and the publicly available online high-resolution imagery World Imagery from ESRI is used as the background. The polygons of the water samples are set to be hollow, displaying only the boundaries. A batch processing program is written to generate a PNG image for each water sample using ArcGIS software.

[0140] (5) Examples of images of different types of samples obtained from classification, such as Figure 3 The example image shown is of a reservoir, such as... Figure 4 The image shown is an example of a pond or dam, such as... Figure 5 The example image shown is of a river, such as... Figure 6 The example image shown is of a lake, such as... Figure 7 The image example shown is an artificial pit, such as... Figure 8 The image example shown is a natural pond, such as... Figure 9 The example image shown is of a ditch, such as... Figure 10 The image shown is an example of a temporary body of water.

[0141] To better understand this invention, specific embodiments have been described in detail above, but these are not intended to limit the invention. Any simple modifications made to the above embodiments based on the technical essence of this invention still fall within the scope of this invention. Each embodiment in this specification focuses on its differences from other embodiments; similar or identical parts between embodiments can be referred to mutually. For system embodiments, since they basically correspond to method embodiments, the description is relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

Claims

1. A method for constructing a sample database for intelligent identification of surface water bodies based on remote sensing imagery, comprising acquiring land use data and satellite remote sensing imagery, characterized in that, It also includes the following steps: Step 1: Extract the surface water body polygons and corresponding water body center points from the land use data. Each water body center point represents a water body. Step 2: Based on the distribution of the water body center points, draw quadrat areas of different sizes. This includes setting quadrat center points at intervals of 5° longitude × 5° latitude, and searching for the water body center points surrounding each quadrat center point from near to far. The longitude and latitude distance when the number of water bodies reaches 100 is used as the quadrat area radius. If the number of water bodies is less than 100 when the quadrat area radius is 5 degrees, then 5 degrees is used as the quadrat area radius. The formula is as follows: ， Among them, R i Let be the radius of quadrat i, and n be the amount of water in the quadrat with a radius of 5. 100 The distance from the center point of the quadrat to the center point of quadrat i is the distance of the water body farthest from the center point of the quadrat. The formula for calculating the distance from the water body to the center point of the quadrat is: ， Among them, l i,j x is the distance from water body j to the center point of sample plot i. i and y i These are the latitude and longitude coordinates of the center point of the quadrat, x and y respectively. j and y j These are the latitude and longitude coordinates of the center point of the water body; Draw a buffer zone for each quadrat area with the center point of each quadrat area as the center and the radius of the quadrat area as the distance, to obtain the circular quadrat areas of all quadrats; Step 3: Select the water body within the quadrat area as the sample. This includes using the Identity tool in ArcGIS software with the circular quadrat area and the center point layer of the water body to extract the quadrat number into the center point layer of the water body. Then, using the join and Calculate Field tools, associate the quadrat number with the water body polygon layer. The water body polygon with the quadrat number is the selected water body sample. Step 4: Generate an image for each sample using the satellite remote sensing imagery, including the following sub-steps: Step 41: Load the sample point layer into ArcGIS software, load the online high-resolution imagery World Imagery published by ESRI as the background, set the water sample polygons to be hollow, and only display the boundaries; Step 42: Write a batch processing program to generate a PNG image for each water sample using ArcGIS software; Step 5: Classify the samples based on the classification features of remote sensing images and the classification features of polygon shapes; Step 6: Create an information table for all samples.

2. The method for constructing a sample database for intelligent identification of surface water bodies based on remote sensing imagery as described in claim 1, characterized in that, Step 1 includes extracting rasters with the attribute of water body from 10m resolution land use raster data, converting the obtained water body rasters into vector polygons using ArcGIS software, and saving them as polygon files in shapehp format; using the FeatureToPoint tool to extract the coordinates of the center point of each polygon, and saving them as point files in shapehp format to represent the geographical location of the corresponding water body polygon.

3. The method for constructing a sample database for intelligent identification of surface water bodies based on remote sensing imagery as described in claim 2, characterized in that, The generation method includes the following sub-steps: Step 421: Obtain each water body polygon object; Step 422: Obtain the bounding box of the water body polygon; Step 423: Enlarge the bounding box of the water body polygon; Step 424: Assign the coordinates of the enlarged water body polygon bounding box to the map; Step 425: Set the image export path; Step 426: Execute the image export command.

4. The method for constructing a sample database for intelligent identification of surface water bodies based on remote sensing imagery as described in claim 3, characterized in that, The types of samples classified include: reservoirs, ponds, rivers, lakes, artificial ponds, natural ponds, ditches, and temporary water bodies.

5. The method for constructing a sample database for intelligent identification of surface water bodies based on remote sensing imagery as described in claim 4, characterized in that, Step 6 includes standardizing the sample results to form a water body classification remote sensing sample library.

6. The method for constructing a sample database for intelligent identification of surface water bodies based on remote sensing imagery as described in claim 5, characterized in that, Step 6 includes the following sub-steps: Step 61: Based on the classification results, establish a sample information table, including sample number, sample type, longitude of the center point, latitude of the center point, and water area; Step 62: Store the samples in folders according to their type and number them sequentially. Reservoirs, ponds, rivers, lakes, artificial ponds, natural ponds, ditches, and temporary water bodies are numbered a, b, c, d, e, f, g, and h respectively, and then numbered sequentially using 6-digit numbers.

Images