Mountainous river roughness coefficient measuring and calculating method based on remote sensing

Abstract

The invention discloses a mountain river course roughness coefficient measuring and calculating method based on remote sensing, and the method comprises the steps: obtaining the geographic space data of a mountain river course, carrying out the data processing of the geographic space data, building a hydrodynamic model calculation grid based on the geographic space processing data, and calculating the roughness coefficient of the mountain river course. And performing roughness assignment on the hydrodynamic model calculation unit of the dry-season river channel based on the remote sensing image data, performing roughness assignment on the hydrodynamic model calculation unit of the non-dry-season river channel, and correcting the roughness value of the hydrodynamic calculation unit to obtain the corrected roughness value of the hydrodynamic calculation unit. According to the method, remote sensing image features in dry seasons and non-dry seasons are extracted and recognized through the remote sensing technology, the roughness value of each calculation unit is calculated according to the size of the calculation unit of the hydrodynamic model, and therefore distributed assignment of calculation unit division of the hydrodynamic model is achieved. According to the method, the actual roughness coefficient can be accurately reflected, so that the simulation precision and reliability of the hydrodynamic model are improved.

Description

Technical Field

[0001] The invention belongs to the technical field of water conservancy engineering, and in particular relates to a method for calculating the roughness coefficient of a mountain river based on remote sensing. Background Art

[0002] In the research and application of hydrodynamic models, the accurate determination of the roughness coefficient is crucial for simulating the flow movement of rivers. At present, the existing hydrodynamic models have the following limitations when determining the roughness coefficient: on the one hand, most of the existing hydrodynamic models are assigned based on experience. This empirical assignment method lacks scientificity and accuracy, because different rivers have different terrain, landforms, vegetation coverage and other characteristics, and it is difficult to accurately determine the roughness coefficient based on experience alone. On the other hand, the existing assignment method is usually based on segmented assignment of river sections or the entire river. This method does not take into account the spatial heterogeneity of rivers. Since different sections of the same river may have different underlying surface types and water flow characteristics, if a unified roughness coefficient is used for assignment, the simulation results will be inaccurate. Summary of the invention

[0003] To solve the above problems, the present invention provides a method for calculating the roughness coefficient along a mountainous river based on remote sensing, so as to solve the problem that the existing roughness coefficient calculation method lacks scientificity and accuracy, is difficult to accurately reflect the actual roughness coefficient, and leads to inaccurate simulation results.

[0004] A method for calculating the roughness coefficient of mountainous rivers based on remote sensing, including:

[0005] Obtain geographic spatial data along mountain rivers, including remote sensing image data and DEM data in dry season and non-dry season;

[0006] Performing data processing on the geospatial data to obtain geospatial processed data;

[0007] Creating hydrodynamic model computational grids based on geospatial processing data;

[0008] Assign roughness to the hydrodynamic model calculation unit of the dry season river channel based on remote sensing image data, and

[0009] Based on remote sensing image data, roughness values ​​are assigned to the hydrodynamic model calculation units of the river channel in the non-dry season;

[0010] The roughness value of the hydrodynamic calculation unit is corrected to obtain a corrected roughness value of the hydrodynamic calculation unit, wherein the roughness value of the hydrodynamic calculation unit includes the roughness value of the hydrodynamic calculation unit of the river channel in the dry season and the river channel in the non-dry season.

[0011] According to a specific embodiment of the present invention, data processing is performed on geospatial data to obtain geospatial processed data, including:

[0012] Radiometric calibration and atmospheric correction are performed on remote sensing image data, and river centerlines are generated using geographic information system tools to obtain remote sensing image processing data;

[0013] The Canny edge detection algorithm is used to detect the edges of the DEM data, determine the river valley area and the river channel area, and clip the DEM data of the river valley area and the river channel area to obtain the DEM processed data.

[0014] According to a specific embodiment of the present invention, creating a hydrodynamic model computing grid based on geospatial processing data includes:

[0015] Grid the DEM data in the river channel area and increase the grid density in the river bends, rapids and shallow areas;

[0016] The gridded DEM data is imported into the hydrodynamic model software to generate a hydrodynamic model calculation grid covering the river channel area.

[0017] According to a specific embodiment of the present invention, the roughness assignment of the hydrodynamic model calculation unit of the dry season river channel based on the remote sensing image data includes:

[0018] The gray-level co-occurrence matrix is ​​used to extract texture features from remote sensing image data in the dry season, and multiple water surface texture feature parameters are obtained, including energy, contrast, entropy, homogeneity and correlation. Energy is used to represent the uniformity of texture distribution and the repetition of image grayscale, contrast is used to measure the difference between grayscales, entropy is used to represent the uncertainty of grayscale distribution, homogeneity is used to measure the smoothness of texture, and correlation is used to represent the grayscale correlation between pixels.

[0019] The roughness value is determined according to multiple water surface texture characteristic parameters, and the texture roughness assignment matrix T is constructed in combination with the corresponding water surface texture characteristic values. The texture roughness assignment matrix T is:

[0020]

[0021] In the formula, E ij represents the energy of the i-th row and j-th column, C ij represents the contrast of the i-th row and j-th column, F ij represents the entropy of the i-th row and j-th column, H ij represents the homogeneity of the i-th row and j-th column, R ij represents the correlation of row i and column j;

[0022] Based on the texture roughness assignment matrix T, calculate the roughness value contribution weight W of the feature parameter in the i-th row and j-th column ij, roughness value contribution weight W ij The calculation formula is:

[0023]

[0024] In the formula, S ij represents the contribution area of ​​the roughness value corresponding to the texture feature parameter of the i-th row and j-th column, and k and m both represent the roughness value;

[0025] in,

[0026]

[0027] Where, T ij Assign the element in the i-th row and j-th column of the texture roughness matrix to O j is the coverage area of ​​characteristic parameter j, S total is the total area covered by remote sensing images;

[0028] Based on the texture roughness assignment matrix T and the roughness value contribution weight W ij Determine the roughness value A of the calculation unit of the hydrodynamic model of the river channel in the dry season i , roughness value A i The calculation formula is:

[0029]

[0030] Where n is the number of characteristic parameters.

[0031] According to a specific embodiment of the present invention, the gray level co-occurrence matrix is ​​used to extract texture features from remote sensing image data in the dry season, and the plurality of water surface texture feature parameters are obtained, further comprising:

[0032] Convert color remote sensing images into grayscale images and determine the grayscale levels of the grayscale images;

[0033] Count the frequency of occurrence of gray value pairs of each pixel in the gray image at distance d and direction θ to form a gray level co-occurrence matrix P(i, j, d, θ);

[0034] Five characteristic parameters, namely energy, contrast, entropy, homogeneity and correlation, are extracted from the gray-level co-occurrence matrix P(i, j, d, θ) and the corresponding eigenvalues ​​are calculated.

[0035] The formula for calculating the eigenvalue of energy is:

[0036]

[0037] Where U is the number of gray levels, and P(i,j) is the element value of the i-th row and j-th column in the gray-level co-occurrence matrix;

[0038] The formula for calculating the eigenvalue of contrast is:

[0039]

[0040] The formula for calculating the eigenvalue of entropy is:

[0041]

[0042] The formula for calculating the eigenvalue of homogeneity is:

[0043]

[0044] The formula for calculating the eigenvalue of correlation is:

[0045]

[0046] in,

[0047]

[0048] In the formula, u x is the mean gray level in the x direction, P(i) is the gray level probability distribution of the i-th row;

[0049]

[0050] In the formula, u y is the mean gray level in the y direction, P(j) is the gray level probability distribution of the jth column;

[0051]

[0052] In the formula, is the variance of the gray level in the x direction;

[0053]

[0054] In the formula, is the variance of the gray level in the y direction.

[0055] According to a specific embodiment of the present invention, the roughness assignment of the hydrodynamic model calculation unit of the non-dry season river channel based on the remote sensing image data includes:

[0056] The supervised classification method is used to classify the features of the remote sensing images of the non-dry season river channel to obtain the feature data;

[0057] A roughness assignment coding rule table is compiled based on the terrain feature data, and roughness is assigned to the hydrodynamic model calculation unit according to the roughness assignment coding rule table.

[0058] According to a specific embodiment of the present invention, the supervised classification method is used to classify the ground feature of the remote sensing image of the non-dry season river channel, and the ground feature data obtained further includes:

[0059] According to the river channel roughness value table, the river channel types and their branches are classified and coded, and a river channel category coding rule table is generated;

[0060] The normalized water index method is used to identify the water body and ground feature characteristics of the non-dry season river remote sensing image, and the corresponding ground feature category code is marked on the remote sensing image according to the river channel category coding rule table;

[0061] The gray level co-occurrence matrix is ​​used to extract the features of the encoded remote sensing image, and the energy and contrast are selected as the feature data of the object.

[0062] According to a specific embodiment of the present invention, compiling a roughness assignment coding rule table based on the feature data, and assigning roughness to the hydrodynamic model calculation unit according to the roughness assignment coding rule table further includes:

[0063] The texture roughness of the remote sensing image is divided based on energy and contrast to obtain a texture-roughness degree coding table, and the roughness assignment coding rule table is generated in combination with the river channel category coding rule table;

[0064] According to the roughness assignment coding rule table, a texture-roughness assignment matrix is ​​constructed and a roughness vector file is generated. The roughness vector file is embedded with the hydrodynamic model calculation grid, and distributed roughness assignment is performed for the hydrodynamic model calculation units of the non-dry season river channel.

[0065] According to a specific embodiment of the present invention, constructing a texture-roughness assignment matrix and generating a roughness vector file according to a roughness assignment coding rule table, embedding the roughness vector file with a hydrodynamic model calculation grid, and performing distributed roughness assignment for a hydrodynamic model calculation unit further comprises:

[0066] Construct the grid cell roughness matrix R according to the roughness assignment coding rule table:

[0067]

[0068] Where RMN represents the roughness value of the grid cell in the Mth row and the Nth column;

[0069] Determine the grid cell roughness vector R′ according to the grid cell roughness matrix R:

[0070] R′=(R 11 R 12 …R 1N R 21 R 22 …R 2N …R M1 RM2 …R MN )

[0071] The roughness vector file is generated according to the grid unit roughness matrix R, and is embedded with the hydrodynamic model calculation grid. The overlap area ratio matrix A with the network calculation unit is constructed according to the grid unit roughness vector R′ in the vector file:

[0072]

[0073] In the formula, A kl represents the overlapping area ratio between the kth computational unit and the lth grid unit;

[0074] According to the grid unit roughness vector R′ and the overlap area ratio matrix A, a comprehensive roughness value matrix S is constructed: S = A·(R′) T =

[0075] S = A·(R′) T =(S1 S2 … S k S k+1 )

[0076] In the formula, S k Represents the comprehensive roughness value of the kth calculation unit.

[0077] According to a specific embodiment of the present invention, the roughness value of the hydrodynamic calculation unit is corrected to obtain the corrected roughness value of the hydrodynamic calculation unit, which includes:

[0078] The roughness value of the hydrodynamic calculation unit is corrected according to the river meandering, that is:

[0079] When the tortuosity is <1.2, keep the roughness value unchanged;

[0080] When 1.2≤ sinuosity<1.5, increase the roughness value by 0.003;

[0081] When the tortuosity is ≥1.5, increase the roughness value by 0.005.

[0082] Compared with the prior art, the present invention has the following advantages:

[0083] The present invention provides a method for calculating the roughness coefficient of a mountain river along the river based on remote sensing. The remote sensing technology is used to extract and identify the remote sensing image features in the dry season and the non-dry season respectively. Then, according to the calculation unit size of the hydrodynamic model, the roughness value of each calculation unit is calculated, thereby realizing distributed assignment based on the division of the hydrodynamic model calculation unit. Compared with the traditional overall or segmented assignment of rivers, the present invention takes into account the seasonal differences and spatial heterogeneity of rivers. According to the different terrain, landform, vegetation coverage and other characteristics of rivers in different seasons, feature extraction and roughness assignment calculation are performed respectively. Therefore, the present invention can accurately reflect the actual roughness coefficient, thereby improving the simulation accuracy and reliability of the hydrodynamic model. BRIEF DESCRIPTION OF THE DRAWINGS

[0084] In order to more clearly illustrate the embodiments of the present disclosure or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present disclosure. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0085] Figure 1 The present invention is a flowchart of a method for calculating the roughness coefficient of a mountain river based on remote sensing according to an embodiment of the present invention.

[0086] Figure 2 It is a flow chart of a method for processing geographic spatial data provided according to an embodiment of the present invention.

[0087] Figure 3 It is a flow chart of a method for creating a hydrodynamic model calculation grid according to an embodiment of the present invention.

[0088] Figure 4 The present invention is a flow chart of a method for assigning roughness to a hydrodynamic model calculation unit of a dry season river channel according to an embodiment of the present invention.

[0089] Figure 5 The present invention is a flowchart of a method for extracting texture features from remote sensing image data in the dry season according to an embodiment of the present invention.

[0090] Figure 6 The present invention is a flow chart of a method for assigning roughness to a hydrodynamic model calculation unit of a non-dry season river channel according to an embodiment of the present invention.

[0091] Figure 7 It is a flow chart of a method for extracting features of land objects from remote sensing images of non-dry season rivers provided according to an embodiment of the present invention.

[0092] Figure 8It is a flow chart of a method for compiling a roughness assignment coding rule table and performing roughness assignment on a hydrodynamic model calculation unit according to the roughness assignment coding rule table provided in accordance with an embodiment of the present invention.

[0093] Fig. 9 It is a flow chart of a method for constructing a texture-roughness assignment matrix and performing roughness assignment on a hydrodynamic model calculation unit according to the texture-roughness assignment matrix provided in accordance with an embodiment of the present invention. DETAILED DESCRIPTION

[0094] In order to make those skilled in the art understand the concept and thought of the present invention more clearly, the present invention is described in detail below in conjunction with specific embodiment.It should be understood that the embodiment provided herein is only a part of all embodiments that the present invention may have.Those skilled in the art, after reading the specification of the application, have the ability to make improvements, transformations, or replacements to part or the whole of the following embodiments, and these improvements, transformations, or replacements are also included in the scope of the present invention.

[0095] In this article, the terms "first", "second" and other similar words are not intended to imply any order, quantity and importance, but are only used to distinguish different elements. In this article, the terms "one", "an" and other similar words are not intended to indicate that there is only one thing, but to indicate that the relevant description is only for one of the things, and the thing may have one or more. In this article, the terms "comprise", "include" and other similar words are intended to indicate logical mutual relations, and cannot be regarded as indicating spatial structural relations. For example, "A includes B" is intended to indicate that B belongs to A logically, but does not mean that B is located inside A in space. In addition, the meanings of the terms "comprise", "include" and other similar words should be regarded as open, rather than closed. For example, "A includes B" is intended to indicate that B belongs to A, but B does not necessarily constitute the whole of A, and A may also include other elements such as C, D, and E.

[0096] In this document, the terms "embodiment", "this embodiment", "one embodiment", and "an embodiment" do not mean that the relevant description is only applicable to a specific embodiment, but rather that the description may also be applicable to one or more other embodiments. Those skilled in the art should understand that in this document, any description of a certain embodiment can be replaced, combined, or combined in other ways with the relevant descriptions in one or more other embodiments, and the new embodiments generated by the replacement, combination, or combination in other ways are easily conceivable by those skilled in the art and fall within the scope of protection of the present invention.

[0097] Example 1

[0098] Additional aspects and advantages of embodiments of the present invention will be described in part in the following description, and will become apparent from the following description, or will be learned through practice of embodiments of the present invention. Figure 1-Figure 9 The embodiment of the present invention provides a method for calculating the roughness coefficient of a mountain river based on remote sensing, comprising:

[0099] S1: Obtain geographic spatial data along mountain rivers, including remote sensing image data and DEM data in the dry season and non-dry season.

[0100] S2: Process the geospatial data to obtain geospatial processed data.

[0101] S3: Create a hydrodynamic model computational grid based on geospatial processing data.

[0102] S4: Assign roughness to the calculation units of the hydrodynamic model of the dry season river channel based on remote sensing image data.

[0103] S5: Assign roughness to the calculation units of the hydrodynamic model of the riverbed in the non-dry season based on remote sensing image data.

[0104] S6: Correcting the roughness value of the hydrodynamic calculation unit to obtain a corrected roughness value of the hydrodynamic calculation unit, wherein the roughness value of the hydrodynamic calculation unit includes the roughness value of the hydrodynamic calculation unit of the river channel in the dry season and the river channel in the non-dry season.

[0105] Specifically, step S1 obtains geographic spatial data along the mountain river, including remote sensing image data and DEM data in the dry season and non-dry season.

[0106] The embodiment of the present invention collects high-resolution optical remote sensing image data of the study area in four seasons, covering a variety of satellite data sources, including but not limited to Landsat series, Sentinel series, Keyhole series, and Gaofen series satellite remote sensing images. Publicly available global digital elevation model DEM data is collected.

[0107] Specifically, step S2 processes the geospatial data to obtain geospatial processed data including:

[0108] S21: Perform radiometric calibration and atmospheric correction on the remote sensing image data, and use geographic information system tools to generate the river centerline to obtain remote sensing image processing data.

[0109] The embodiment of the present invention uses a remote sensing image processing tool to perform radiometric calibration on the remote sensing image data, and uses radiometric calibration technology to correct the pixel value of the remote sensing image to the actual reflectivity or brightness. Then, the remote sensing image data is atmospherically corrected to eliminate the influence of atmospheric scattering and absorption on the image, ensuring that the spectral information accurately reflects the actual surface conditions. The water body index method NDWI is used to extract the water surface vector, and the river centerline is generated based on the water surface vector with the help of a geographic information system tool. The river centerline is used to determine the range of the river valley and the range of the river channel.

[0110] S22: Use the Canny edge detection algorithm to perform edge detection on the DEM data, determine the river valley area and the river channel area, and clip the DEM data of the river valley area and the river channel area to obtain DEM processing data.

[0111] The embodiment of the present invention first creates a buffer zone that is twice the width of the river channel based on the center line of the river channel, and preliminarily delineates the scope of the river valley. Then, the Canny edge detection algorithm is used to detect the edge of the buffer zone, the DEM data of the buffer zone is clipped, and its slope is calculated. Then, a Gaussian filter is applied to smooth the DEM-slope data to eliminate noise, and then the Sobel operator is used to calculate the horizontal and vertical gradients, perform non-maximum suppression, and retain the local maximum point. The edge of the buffer zone is then extracted, and by setting high and low thresholds, strong boundaries and non-boundaries are distinguished, and a complete edge line is formed by edge connection, and finally the edge range of the river valley and the river channel is determined. After extracting the edge, the DEM data of the river valley area and the river channel area are clipped. According to the scope of the river valley and the river channel, the DEM data of the river valley area and the river channel area are clipped to prepare basic terrain data for establishing a grid for the hydrodynamic model.

[0112] Specifically, step S3 of creating a hydrodynamic model calculation grid based on geospatial processing data includes:

[0113] S31: Grid the DEM data in the river channel area and increase the grid density in the river bends, rapids and shallow areas.

[0114] The embodiment of the present invention grids the DEM data of the river channel area according to the river channel range determined in step S2, and selects a suitable grid size according to the complexity of the actual terrain and computing resources. For example, a 10m×10m grid can be selected for flat terrain, and a 20m×20m grid can be selected for complex terrain.

[0115] S32: Import the gridded DEM data into the hydrodynamic model software to generate a hydrodynamic model calculation grid covering the river channel area.

[0116] The embodiment of the present invention imports the DEM data of the river channel area after Canny edge detection and clipping into the hydrodynamic model software to ensure the accurate matching of the DEM data with the river channel area. The grid density is increased in complex areas such as river bends, rapids and shoals to improve the accuracy of the model, and finally a hydrodynamic model calculation grid covering the entire river channel area is generated.

[0117] Specifically, step S4 assigns roughness values ​​to the hydrodynamic model calculation units of the dry season river channel based on the remote sensing image data, including:

[0118] S41: Use gray level co-occurrence matrix to extract texture features from remote sensing image data in dry season, and obtain multiple water surface texture feature parameters, including energy, contrast, entropy, homogeneity and correlation, where energy is used to represent the uniformity of texture distribution and the repetition degree of image grayscale, contrast is used to measure the difference between grayscales and reflect the roughness of texture, entropy is used to represent the uncertainty of grayscale distribution and reflect the complexity of texture, homogeneity is used to measure the smoothness of texture, and correlation is used to represent the grayscale correlation between pixels, specifically including the following steps:

[0119] S411: Convert the color remote sensing image into a grayscale image and determine the grayscale levels of the grayscale image to reduce the amount of calculation.

[0120] S412: Count the frequency of occurrence of gray value pairs at distance d and direction θ for each pixel in the gray image to form a gray level co-occurrence matrix P (i, j, d, θ).

[0121] S413: Extract five characteristic parameters of energy, contrast, entropy, homogeneity and correlation from the gray-level co-occurrence matrix P(i, j, d, θ) respectively, and calculate the corresponding eigenvalues.

[0122] In the dry season, the landforms within the river surface are relatively simple, so the roughness coefficient can be determined by analyzing the texture characteristics of the water surface. If the water surface has obvious texture characteristics, it indicates that there are more gravel, pebbles or large boulders on the river bottom. The roughness coefficient of this type of river section is relatively high, for example, the roughness of large boulders is 0.07, the roughness of pebbles is 0.05, and the roughness of gravel is 0.04. If there are no obvious texture characteristics on the water surface, it indicates that the river bottom is relatively smooth, usually bedrock or stone riverbed, and the roughness coefficient of this type of river section is relatively low, for example, the roughness of bedrock is 0.03. The embodiment of the present invention uses a grayscale co-occurrence matrix to extract texture features from remote sensing image data in the dry season, and extracts five feature parameters of energy, contrast, entropy, homogeneity and correlation respectively. The energy represents the uniformity of the grayscale distribution of the image. The greater the energy, the more regular and uniform the texture pattern. The value range of energy is: 0-1. The higher the energy, the more uniform the image grayscale, the smoother the texture, and the lower the roughness. The lower the energy, the more uneven the grayscale distribution, the more complex the texture, and the greater the roughness. Contrast indicates the strength of the contrast of texture primitives. The higher the contrast, the greater the image roughness. The contrast value range is 0 to infinity. The higher the contrast, the more obvious the grayscale difference, the rougher the texture, and the greater the roughness. The lower the contrast, the less obvious the grayscale change, the smoother the image, and the lower the roughness. Entropy indicates the uncertainty and degree of disorder of the grayscale distribution, and is an indicator to measure the complexity of the image. The entropy value range is greater than 0. The larger the entropy, the more complex the image, the more chaotic the texture, and the greater the roughness. The smaller the entropy, the more ordered the image, the simpler the texture, and the lower the roughness. Homogeneity is used to measure the similarity of pixel pairs in an image. The higher the value, the smoother the texture of the image. The homogeneity value range is 0 to 1. The higher the homogeneity, the smaller the pixel difference, the smoother the image, and the lower the roughness. The lower the homogeneity, the larger the pixel difference, the complex texture, and the greater the roughness. Correlation reflects the linear relationship between pixels in an image, and the correlation value range is -1 to 1. The higher the correlation, the stronger the linear relationship between the grayscale pixels in the image, the more ordered the texture, and the lower the roughness; the lower the correlation, the weaker the relationship between the pixels, the more disordered the image, and the greater the roughness.

[0123] Among them, the energy E ij The formula for calculating the eigenvalue of is:

[0124]

[0125] Where N is the number of gray levels, and P(i, j) is the element in the i-th row and j-th column of the gray-level co-occurrence matrix.

[0126] Contrast C ij The formula for calculating the eigenvalue of is:

[0127]

[0128] Where P(i, j) is the element in the i-th row and j-th column of the gray-level co-occurrence matrix, and ij represents the difference between gray levels.

[0129] Entropy F ij The formula for calculating the eigenvalue of is:

[0130]

[0131] Homogeneity ij The formula for calculating the eigenvalue of is:

[0132]

[0133] Correlation R ij The formula for calculating the eigenvalue of is:

[0134]

[0135] in,

[0136]

[0137] In the formula, u x is the mean value of the gray level in the x direction, P(i) is the gray level probability distribution of the i-th row,

[0138]

[0139] In the formula, u y is the mean value of the gray level in the y direction, P(j) is the gray probability distribution of the jth column,

[0140]

[0141] In the formula, is the variance of the gray level in the x direction.

[0142]

[0143] In the formula, is the variance of the gray level in the y direction.

[0144] S42: Determine the roughness value according to the multiple water surface texture characteristic parameters, and construct a texture roughness assignment matrix T in combination with the corresponding water surface texture characteristic values. The texture roughness assignment matrix T is:

[0145]

[0146] In the formula, E ij represents the energy of the i-th row and j-th column, C ij represents the contrast of the i-th row and j-th column, F ij represents the entropy of the i-th row and j-th column, H ijrepresents the homogeneity of the i-th row and j-th column, R ij represents the correlation of the i-th row and j-th column.

[0147] Table 1 is a table of water surface texture features and roughness values. As shown in Table 1, each row represents a water surface texture feature, from top to bottom, energy E ij , Contrast C ij , Entropy F ij Homogeneity ij and correlation R ij , each column represents a corresponding roughness value. The embodiment of the present invention can determine the corresponding roughness value according to the value ranges of five different water surface texture features.

[0148] Table 1 Water surface texture characteristics and roughness values

[0149]

[0150] S43: Calculate the roughness value contribution weight W of the feature parameter in the i-th row and j-th column based on the texture roughness assignment matrix T. ij , roughness value contribution weight W ij The calculation formula is:

[0151]

[0152] In the formula, S ij represents the contribution area of ​​the roughness value corresponding to the texture feature parameter in the i-th row and j-th column, and k and m both represent the roughness value.

[0153]

[0154] Where, T ij Assign the element in the i-th row and j-th column of the texture roughness matrix to O j is the coverage area of ​​characteristic parameter j, S total is the total area covered by the remote sensing image.

[0155] S44: Based on the texture roughness assignment matrix T and the roughness value contribution weight W oj Determine the roughness value A of the calculation unit of the hydrodynamic model of the river channel in the dry season o , roughness value A o The calculation formula is:

[0156]

[0157] Where n is the number of characteristic parameters.

[0158] Specifically, step S5 assigns roughness values ​​to the hydrodynamic model calculation units of the non-dry season river channel based on the remote sensing image data, including:

[0159] S51: Use supervised classification method to classify the remote sensing images of non-dry season rivers and obtain the feature data.

[0160] In the non-dry season river channel, due to the presence of shallows, vegetation and other coverage, the embodiment of the present invention adopts a supervised classification method to classify the ground objects, and then determine their texture feature parameters. The supervised classification establishes a classification model through pre-labeled training samples. In terms of sample selection, it is necessary to select a certain number of training areas for each category on the image. The selected training samples should be able to accurately represent the characteristics of each category in the entire area, and then calculate the feature information of the training samples of each category, and then select a suitable classification algorithm to classify the image according to the characteristics of the image and classification requirements. Preferably, the embodiment of the present invention adopts a support vector machine (SVM) as a supervised classification algorithm, and classifies samples of different categories by finding the optimal hyperplane. Before classification, the data needs to be preprocessed, which specifically includes the following steps:

[0161] Step 1: Data standardization. To ensure data stability, Z-score standardization is used to unify the eigenvalues ​​to the same scale. The standardization calculation formula is:

[0162]

[0163] Where X is the original eigenvalue, μ is the mean of the feature, and σ is the standard deviation of the feature.

[0164] Step 2: Parameter selection. The performance of SVM depends on the kernel function and penalty parameters. Common kernel functions include linear kernel, polynomial kernel and radial basis function (RBF) kernel. First, the optimal kernel function and penalty parameters are selected through k-fold cross validation. Then, the range of C value and γ value is preset, and the best parameter combination is found through grid search.

[0165] Step 3: Model training. Use the labeled training samples to train the SVM model, adjust the model parameters to minimize the classification error rate, and finally obtain the classification model.

[0166] Step S51 specifically includes the following steps:

[0167] S511: Classify and assign codes to the river channel types and their branches according to the river channel roughness value table, and generate a river channel category coding rule table.

[0168] The embodiment of the present invention refers to the roughness lookup table and determines the roughness value of the natural river channel, and divides the river channel types and characteristics into: small rivers, large rivers, and flood-time floodplains, and further assigns coding rules to its branches to generate a river channel category coding rule table. As shown in Table 1, the coding rule of the river channel category coding rule table is xxx, and the first x is the code for small rivers, large rivers, and flood-time floodplains. The code of the small river is set to 1, the code of the large river is set to 2, and the code of the flood-time floodplain is set to 3. The second x is the branch of the first x. For the small river type, there are two branches: plain rivers and mountain rivers. The code of the plain river is set to 1, and the code of the mountain river is set to 2. For flood-time floodplains, there are 4 branches, namely: grassland, no bushes; cultivated land; bushes; trees. The third x is the branch under the second x.

[0169] Table 1

[0170]

[0171]

[0172] S512: Use the normalized water index method to identify the water body and ground feature characteristics of the non-dry season river remote sensing image, and mark the corresponding ground feature category code on the remote sensing image according to the river channel category coding rule table.

[0173] The Normalized Difference Water Index (NDWI) is a remote sensing index used to identify and map surface water bodies. It highlights water information by comparing the reflectance of the green band and the near-infrared band in remote sensing images. The value threshold of the Normalized Difference Water Index (NDWI) is usually between -1 and 1, and positive values ​​indicate a higher probability of water bodies. The calculation formula of the Normalized Difference Water Index (NDWI) is:

[0174]

[0175] Where Green is the reflectivity of the green light band, and NIR is the reflectivity of the near infrared band.

[0176] S513: Extract the ground feature of the encoded remote sensing image using the gray level co-occurrence matrix, and select energy and contrast as ground feature data. The energy reflects the uniformity of the gray distribution of the image. The greater the energy, the more regular and uniform the texture pattern. The contrast indicates the strength of the contrast of the texture primitives. The higher the contrast, the greater the image roughness.

[0177] S52: compiling a roughness assignment coding rule table based on the feature data, and performing roughness assignment on the hydrodynamic model calculation unit according to the roughness assignment coding rule table, specifically including the following steps:

[0178] S521: The texture roughness of the remote sensing image is divided based on energy and contrast to obtain a texture-roughness degree coding table, and a roughness assignment coding rule table is generated in combination with the river channel category coding rule table.

[0179] The embodiment of the present invention is based on the energy E ij and contrast C ij The calculated value of is used to divide the texture roughness of the remote sensing image. The division criteria are shown in Table 2. When the energy is high (close to 1) and the contrast is low, the roughness takes the minimum value and is encoded as 1. When the energy and contrast are both at the mean, the roughness takes the normal value and is encoded as 2. When the energy is low and the contrast is high, the roughness takes the maximum value and is encoded as 3.

[0180] Table 2

[0181]

[0182] Table 3 is a roughness value table. Combined with the classification standard in Table 2, the embodiment of the present invention determines the specific roughness value according to the description in Table 3.

[0183] Table 3

[0184]

[0185]

[0186] Combined with the river channel category coding rule table, the roughness assignment coding rule table shown in Table 4 is generated:

[0187] Table 4

[0188]

[0189] S522: constructing a texture-roughness assignment matrix according to the roughness assignment coding rule table and generating a roughness vector file, embedding the roughness vector file with the hydrodynamic model calculation grid, and performing distributed roughness assignment for the hydrodynamic model calculation unit of the non-dry season river channel, specifically including:

[0190] S5221: Construct the grid unit roughness matrix R according to the roughness assignment coding rule table:

[0191]

[0192] Where RMN represents the roughness value of the grid cell in the Mth row and the Nth column.

[0193] The embodiment of the present invention assigns a roughness value to each grid cell on the remote sensing image according to the roughness assignment coding rule table, and arranges the roughness value of each grid cell in an M×N matrix to form an M×N grid cell roughness matrix R.

[0194] S5222: Determine the grid cell roughness vector R′ according to the grid cell roughness matrix R:

[0195] R′=(R 11 R 12 …R 1N R 21 R 22 …R 2N …R M1 R M2 …R MN ) (17)

[0196] The present invention implements the method of expanding the M×N grid cell roughness matrix R into a (M×N)×1 column vector R′, where each element RMN in R′ represents the roughness value of the grid cell in the Mth row and the Nth column.

[0197] S5223: Generate a roughness vector file according to the grid unit roughness matrix R, and fit it with the hydrodynamic model calculation grid. Construct an overlapping area ratio matrix A with the grid calculation unit according to the grid unit roughness vector R′ in the vector file:

[0198]

[0199] In the formula, A kl Represents the overlapping area ratio of the kth computational unit to the lth grid unit.

[0200] In the embodiment of the present invention, the overlapping area ratios of the calculation units of the hydrodynamic model and the generated vector files form an overlapping area ratio matrix A of P×(M×N), where P is the number of the hydrodynamic model calculation units, M×N is the number of the vector file units, and each A of the matrix A is kl Represents the overlapping area ratio of the kth computation unit to the lth vector file unit.

[0201] S5224: Construct a comprehensive roughness value matrix S according to the grid unit roughness vector R′ and the overlap area ratio matrix A:

[0202] S = A·(R′) T =(S1 S2 … S k S k+1 ) (19)

[0203] In the formula, S k Represents the comprehensive roughness value of the kth calculation unit.

[0204] The embodiment of the present invention uses matrix multiplication S=A·(R′) T , calculate the comprehensive roughness value of each calculation unit. Where S represents the comprehensive roughness value of each calculation unit.

[0205] Specifically, step S6 corrects the roughness value of the hydrodynamic calculation unit to obtain the corrected roughness value of the hydrodynamic calculation unit, including:

[0206] The roughness value of the hydrodynamic calculation unit is corrected according to the river meandering, that is:

[0207] When the tortuosity is <1.2, keep the roughness value unchanged.

[0208] When 1.2≤tortuosity<1.5, increase the roughness value by 0.003.

[0209] When the tortuosity is ≥1.5, increase the roughness value by 0.005.

[0210] Since the channel curvature has a great influence on the roughness coefficient, the embodiment of the present invention further introduces the channel sinuosity to judge the straightness and curvature of the channel, so as to automatically correct the roughness assignment of the hydrodynamic model.

[0211] The river meandering is used to determine the degree of curvature of the river, and is usually defined as the ratio of the actual length of the river to the straight-line distance between the starting point and the end point of the river. The specific calculation formula is as follows:

[0212]

[0213] Where L is the actual length of the river, that is, the total length measured by the centerline of the river, and D is the straight-line distance between the starting point and the end point of the river.

[0214] The roughness value of the hydrodynamic calculation unit is corrected by the river meandering, making the calculated roughness value more accurate.

[0215] In summary, the method for calculating the roughness coefficient of a mountain river based on remote sensing according to the present invention has the following advantages:

[0216] The present invention provides a method for calculating the roughness coefficient of a mountain river along the river based on remote sensing. The remote sensing technology is used to extract and identify the remote sensing image features in the dry season and the non-dry season respectively. Then, according to the calculation unit size of the hydrodynamic model, the roughness value of each calculation unit is calculated, thereby realizing distributed assignment based on the division of the hydrodynamic model calculation unit. Compared with the traditional overall or segmented assignment of rivers, the present invention takes into account the seasonal differences and spatial heterogeneity of rivers. According to the different terrain, landform, vegetation coverage and other characteristics of rivers in different seasons, feature extraction and roughness assignment calculation are performed respectively. Therefore, the present invention can accurately reflect the actual roughness coefficient, thereby improving the simulation accuracy and reliability of the hydrodynamic model.

[0217] The above describes in detail the concepts, principles and ideas of the present invention in conjunction with specific implementation methods (including embodiments and examples). Those skilled in the art should understand that the implementation methods of the present invention are not limited to the forms given above. After reading this application document, those skilled in the art can make any possible improvements, substitutions and equivalent forms to the steps, methods, systems and components in the above implementation methods. These improvements, substitutions and equivalent forms should be deemed to fall within the scope of the present invention, and the protection scope of the present invention is subject only to the claims.

Claims

1. A method for calculating the roughness coefficient of a mountain river based on remote sensing, characterized in that: include: Obtain geographic spatial data along mountain rivers, including remote sensing image data and DEM data in dry season and non-dry season; Performing data processing on the geospatial data to obtain geospatial processed data; creating a hydrodynamic model computational grid based on the geospatial processed data; Based on the remote sensing image data, roughness values ​​are assigned to the hydrodynamic model calculation units of the dry season river channel, and Based on the remote sensing image data, roughness values ​​are assigned to the hydrodynamic model calculation units of the non-dry season river channel; The roughness value of the hydrodynamic calculation unit is corrected to obtain a corrected roughness value of the hydrodynamic calculation unit, wherein the roughness value of the hydrodynamic calculation unit includes the roughness values ​​of the hydrodynamic calculation unit of the river channel in the dry season and the river channel in the non-dry season.

2. The method for calculating the roughness coefficient of a mountain river based on remote sensing according to claim 1 is characterized in that: The performing data processing on the geospatial data to obtain geospatial processed data comprises: Performing radiometric calibration and atmospheric correction on the remote sensing image data, and generating a river centerline using a geographic information system tool to obtain remote sensing image processing data; The Canny edge detection algorithm is used to perform edge detection on the DEM data to determine the river valley area and the river channel area, and the DEM data of the river valley area and the river channel area are clipped to obtain DEM processed data.

3. The method for calculating the roughness coefficient of a mountain river based on remote sensing according to claim 2 is characterized in that: The creating of a hydrodynamic model computational grid based on the geospatial processing data comprises: Meshing the DEM data of the river channel area, and increasing the mesh density in the river bends, rapids and shoals; The gridded DEM data is imported into the hydrodynamic model software to generate a hydrodynamic model calculation grid covering the river channel area.

4. The method for calculating the roughness coefficient of a mountain river based on remote sensing according to claim 1, characterized in that: The roughness assignment of the hydrodynamic model calculation unit of the dry season river channel based on the remote sensing image data comprises: The gray-level co-occurrence matrix is ​​used to extract texture features from remote sensing image data in the dry season, and multiple water surface texture feature parameters are obtained, including energy, contrast, entropy, homogeneity and correlation. Energy is used to represent the uniformity of texture distribution and the repetition of image grayscale, contrast is used to measure the difference between grayscales, entropy is used to represent the uncertainty of grayscale distribution, homogeneity is used to measure the smoothness of texture, and correlation is used to represent the grayscale correlation between pixels. The roughness value is determined according to the plurality of water surface texture characteristic parameters, and a texture roughness assignment matrix T is constructed in combination with the corresponding water surface texture characteristic values. The texture roughness assignment matrix T is: In the formula, E ij represents the energy of the i-th row and j-th column, C ij represents the contrast of the i-th row and j-th column, F ij represents the entropy of the i-th row and j-th column, H ij represents the homogeneity of the i-th row and j-th column, R ij represents the correlation of row i and column j; Based on the texture roughness assignment matrix T, the roughness value contribution weight W of the feature parameter in the i-th row and j-th column is calculated. ij , the roughness value contribution weight W ij The calculation formula is: In the formula, S ij represents the contribution area of ​​the roughness value corresponding to the texture feature parameter of the i-th row and j-th column, and k and m both represent the roughness value; in, Where, T ij Assign the element in the i-th row and j-th column of the texture roughness matrix to O j is the coverage area of ​​characteristic parameter j, S total is the total area covered by remote sensing images; Based on the texture roughness assignment matrix T and the roughness value contribution weight W ij Determine the roughness value A of the calculation unit of the hydrodynamic model of the river channel in the dry season i , the roughness value A i The calculation formula is: Where n is the number of characteristic parameters.

5. The method for calculating the roughness coefficient of a mountain river based on remote sensing according to claim 4 is characterized in that: The method of extracting texture features from remote sensing image data in the dry season using a gray level co-occurrence matrix to obtain a plurality of water surface texture feature parameters further includes: Converting a color remote sensing image into a grayscale image and determining the grayscale level of the grayscale image; Counting the frequency of occurrence of gray value pairs of each pixel in the gray image at distance d and direction θ to form a gray level co-occurrence matrix P(i, j, d, θ); Five characteristic parameters, namely energy, contrast, entropy, homogeneity and correlation, are extracted from the gray-level co-occurrence matrix P(i, j, d, θ) and the corresponding eigenvalues ​​are calculated. The energy characteristic value calculation formula is: Where U is the number of gray levels, and P(i,j) is the element value of the i-th row and j-th column in the gray-level co-occurrence matrix; The characteristic value calculation formula of the contrast is: The characteristic value calculation formula of the entropy is: The formula for calculating the characteristic value of homogeneity is: The calculation formula of the characteristic value of the correlation is: in, In the formula, u x is the mean gray level in the x direction, P(i) is the gray level probability distribution of the i-th row; In the formula, u y is the mean gray level in the y direction, P(j) is the gray level probability distribution of the jth column; In the formula, is the variance of the gray level in the x direction; In the formula, is the variance of the gray level in the y direction.

6. The method for calculating the roughness coefficient of a mountain river based on remote sensing according to claim 1, characterized in that: The roughness assignment of the hydrodynamic model calculation unit of the non-dry season river channel based on the remote sensing image data comprises: The supervised classification method is used to classify the features of the remote sensing images of the non-dry season river channel to obtain the feature data; A roughness assignment coding rule table is compiled based on the land feature feature data, and roughness is assigned to a hydrodynamic model calculation unit according to the roughness assignment coding rule table.

7. The method for calculating the roughness coefficient of a mountain river based on remote sensing according to claim 6, characterized in that: The method of using the supervised classification method to classify the features of the remote sensing images of the non-dry season river to obtain the feature data further includes: According to the river channel roughness value table, the river channel types and their branches are classified and coded, and a river channel category coding rule table is generated; The normalized water index method is used to identify the water body and ground feature characteristics of the non-dry season river channel remote sensing image, and the corresponding ground feature category code is marked on the remote sensing image according to the river channel category coding rule table; The gray level co-occurrence matrix is ​​used to extract the features of the objects in the encoded remote sensing image, and the energy and contrast are selected as the feature data of the objects.

8. The method for calculating the roughness coefficient of a mountain river based on remote sensing according to claim 7, characterized in that: The step of compiling a roughness assignment coding rule table based on the land feature data and assigning roughness to the hydrodynamic model calculation unit according to the roughness assignment coding rule table further includes: The texture roughness of the remote sensing image is divided based on the energy and the contrast to obtain a texture-roughness degree coding table, and a roughness assignment coding rule table is generated in combination with the river channel category coding rule table; A texture-roughness assignment matrix is ​​constructed according to the roughness assignment coding rule table and a roughness vector file is generated. The roughness vector file is embedded with a hydrodynamic model calculation grid, and distributed roughness assignment is performed for a hydrodynamic model calculation unit of a non-dry season river channel.

9. The method for calculating the roughness coefficient of a mountain river based on remote sensing according to claim 8, characterized in that: The step of constructing a texture-roughness assignment matrix according to the roughness assignment coding rule table and generating a roughness vector file, embedding the roughness vector file with a hydrodynamic model calculation grid, and performing distributed roughness assignment for a hydrodynamic model calculation unit further comprises: The grid unit roughness matrix R is constructed according to the roughness assignment coding rule table: Where RMN represents the roughness value of the grid cell in the Mth row and the Nth column; Determine the grid cell roughness vector R′ according to the grid cell roughness matrix R: R′=(R 11 R 12 …R 1N R 21 R 22 …R 2N …R M1 R M2 …R MN ) A roughness vector file is generated according to the grid unit roughness matrix R, and is embedded with the hydrodynamic model calculation grid. An overlap area ratio matrix A with the network calculation unit is constructed according to the grid unit roughness vector R′ in the vector file: In the formula, A kl represents the overlapping area ratio between the kth computational unit and the lth grid unit; The comprehensive roughness value matrix S is constructed according to the grid unit roughness vector R′ and the overlap area ratio matrix A: S=A·(R′) T =(S1 S2…S k S k+1 ) In the formula, S k Represents the comprehensive roughness value of the kth calculation unit.

10. The method for calculating the roughness coefficient of a mountain river based on remote sensing according to claim 1, characterized in that: The step of correcting the roughness value of the hydrodynamic calculation unit to obtain the corrected roughness value of the hydrodynamic calculation unit includes: The roughness value of the hydrodynamic calculation unit is corrected according to the river channel sinuosity, that is: When the tortuosity is <1.2, keep the roughness value unchanged; When 1.2≤ sinuosity<1.5, increase the roughness value by 0.003; When the tortuosity is ≥1.5, increase the roughness value by 0.005.

Images