River channel reach morphological feature recognition method based on remote sensing

By using multi-source remote sensing data fusion technology, river sections are precisely divided and morphological feature parameters are extracted, solving the problem of imprecise river section identification in existing technologies. This enables accurate identification and dynamic monitoring of river morphological features, improving the scientific nature and efficiency of river management.

CN121074667AActive Publication Date: 2025-12-05CHINA CONSTR WATER ENVIRONMENTAL PROTECTION CO LTD

Patent Information

Application Number
CN202511237111.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-09-01
Publication Date
2025-12-05
Estimated Expiration
2045-09-01

AI Technical Summary

Technical Problem

Existing technologies for identifying the morphological features of river sections suffer from problems such as insufficiently precise segmentation, incomplete extraction of morphological feature parameters, and inadequate ability to analyze dynamic change trends. As a result, the evolution patterns of river channels are difficult to accurately reflect, making it difficult to meet management needs.

Method used

Using multi-source remote sensing data fusion technology, the river boundary line and centerline are extracted through optical satellite imagery, synthetic aperture radar imagery and lidar point cloud data. The river is divided into several river segments by combining the curvature change characteristics of the river centerline, morphological feature parameters are extracted and classified, and time series analysis is performed to predict the river morphological characteristics at future nodes.

Benefits of technology

It enables precise identification and dynamic monitoring of river morphology characteristics, improves the scientific nature and efficiency of river management, provides refined data support for river management, supports intelligent and data-driven decision-making in river management, and ensures the reliability and applicability of prediction results.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121074667A_ABST
    Figure CN121074667A_ABST
Patent Text Reader

Abstract

The invention belongs to the technical field of river reach feature recognition, and particularly relates to a river reach morphological feature recognition method based on remote sensing. Through multi-source remote sensing data fusion analysis, accurate recognition and dynamic monitoring of river channel morphological characteristics are achieved, scientificity and efficiency of river channel treatment are improved, refined data support is provided for river channel ecological restoration and management through river reach classification and construction of a morphological module, and the river channel ecological restoration and management method is suitable for river channel ecological restoration and management. In this way, the intelligent and datamation requirements of river regulation decision making are supported, meanwhile, effective prediction of the river form evolution trend is achieved through time sequence analysis of the repair node and the current node, verification and evaluation of the confidence degree are conducted after the prediction result is output, and therefore the reliability and applicability of the prediction result are ensured, and the prediction efficiency is improved. The prediction result can provide corresponding data support for a river regulation project, the river form change trend can be pre-judged in advance, and a scientific basis is provided for flood control and disaster reduction and water resource management.
Need to check novelty before this filing date? Find Prior Art

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of river section feature recognition, and particularly relates to a river section morphological feature recognition method based on remote sensing. BACKGROUND

[0002] With the increasing demand for water resource management, the management of river channels is more refined, and is no longer limited to the overall feature analysis of river channels, but goes deep into the morphological feature recognition of river sections, so as to accurately grasp and scientifically predict the evolution law of river channels. Traditional river morphological feature recognition methods mostly rely on manual field measurement, and have problems such as large workload, long cycle, limited coverage, etc., and are difficult to meet the monitoring demand of a large range and high frequency. However, with the rapid development of remote sensing technology, a new means is provided for efficient recognition of river section morphological features, which can greatly improve the data collection efficiency, effectively reduce the labor cost, and has strong spatio-temporal coverage capability, so that dynamic monitoring of the evolution process of river channels becomes possible.

[0003] In the prior art, although there are some technical solutions for river feature recognition based on remote sensing data, there are problems such as insufficient division of river sections, incomplete extraction of morphological feature parameters, and insufficient dynamic change trend analysis capability. For example, some technical solutions only divide the river channel into straight sections according to the geometric center line, without fully considering the actual bending characteristics and morphological complexity of the river section, resulting in rough river section classification and difficulty in accurately reflecting the real evolution law of the river channel. In addition, the existing methods are often limited to basic indicators such as width and length in the selection of morphological feature parameters, and lack effective extraction of high-order features such as river section terrain undulation and slope change, which limits the understanding of the evolution process of the river channel, thereby resulting in low prediction accuracy and difficulty in meeting the actual needs of river management and governance. Based on this, the present application provides a river section morphological feature recognition method based on remote sensing to solve the above problems. SUMMARY

[0004] The purpose of the present application is to provide a river section morphological feature recognition method based on remote sensing, which can effectively improve the fineness of river section division and the comprehensiveness of morphological feature parameter extraction, so as to more accurately reflect the evolution law of the river channel.

[0005] The technical scheme adopted by the present application is as follows: A river section morphological feature recognition method based on remote sensing, comprising: obtaining multi-source remote sensing data of a target river channel, wherein the multi-source remote sensing data comprises optical satellite images, synthetic aperture radar images and laser radar point cloud data; cooperatively preprocessing the multi-source remote sensing data to obtain a standardized remote sensing data set; extract a river boundary line and a center line based on multi-source remote sensing data, and divide the river into a plurality of river sections according to a curvature variation feature of the river center line; extract a morphological feature parameter of each river section, including a river section length, a width, a bending degree, and a river bank slope, and then classify each river section according to the morphological feature parameter to obtain a plurality of river section types, and merge the river sections of the same type and adjacent to each other to form a continuous river morphological module; obtain a repair node of each river section, and construct a sample collection period between the repair node and a current node, and perform time series analysis on the morphological feature parameter in the sample collection period to identify a morphological feature variation trend of each river section, and then predict a river morphological feature parameter at a future node based on the morphological feature variation trend.

[0006] In a preferred scheme, the step of obtaining multi-source remote sensing data of the target river includes: a plurality of monitoring points are pre-deployed along the river bank, and each monitoring point includes a collection device for collecting multi-source remote sensing data, the collection device including an optical camera, a synthetic aperture radar antenna, and a laser radar scanner; synchronously obtain optical satellite images, synthetic aperture radar images, and laser radar point cloud data of the target river by starting each collection device at a fixed time, and aggregate the initial remote sensing data; add a time stamp and a geographic label to the aggregated initial remote sensing data to form multi-source remote sensing data with space-time identification.

[0007] In a preferred scheme, the step of synchronously preprocessing the multi-source remote sensing data includes: unify a spatial reference of the optical satellite images, the synthetic aperture radar images, and the laser radar point cloud data, and register the optical satellite images, the synthetic aperture radar images, and the laser radar point cloud data to a unified coordinate system; perform multi-temporal image superposition and band reconstruction on the optical satellite images to recover the hidden area; perform denoising and geometric correction on the synthetic aperture radar images; perform filtering and elevation normalization on the laser radar point cloud data; fuse the preprocessed optical satellite images, synthetic aperture radar images, and laser radar point cloud data to integrate them into a unified information layer to form a standardized remote sensing data set with a unified spatial reference.

[0008] In a preferred scheme, the step of extracting a river boundary line and a center line based on multi-source remote sensing data, and dividing the river into a plurality of river sections according to a curvature variation feature of the river center line includes: The river water mask is calculated based on the spectral characteristics of optical satellite images, and the boundary of the water mask is corrected based on the microwave reflection characteristics of synthetic aperture radar images. The river terrain elevation information is extracted from the laser radar point cloud data, and the river bank line is determined according to the slope change. The river boundary line is extracted according to the fusion of the river water mask and the river bank line information, and the river center line is extracted according to the river boundary line and the terrain elevation data. The curvature of the river center line is calculated, and the curvature peak points are marked. The feature points that exceed twice the standard deviation of the overall curvature fluctuation range and have a distance between adjacent curvature peak points greater than the average width of the river are extracted as river section division nodes. According to the river section division nodes, the river is divided into several independent river sections, and the boundary range of each independent river section is extracted based on the center line information of the river section.

[0009] In a preferred embodiment, the step of extracting the morphological feature parameters of each river section includes: Based on the sequence of coordinates of the center line of each river section, the length of the river section is calculated by accumulating the distance between adjacent coordinate points. A normal profile line is generated along the center line of the river section, and the width of the river section is determined according to the distance between the intersection points of the river boundary line and the normal profile line. The straight-line distance between the starting point and the ending point of the center line of the river section is obtained, and the tortuosity of the river section is calculated based on the actual length of the river section, wherein the value of the tortuosity is the ratio of the actual length of the river section to the straight-line distance between the starting point and the ending point of the center line of the river section. The elevation points of the slope surface on both banks are extracted along the river boundary line, the slope angle of the slope surface is fitted by linear regression, and the average slope of both banks is taken as the slope parameter of the river section. The length, width, tortuosity and slope parameter of the river section are recorded as morphological feature parameters.

[0010] In a preferred embodiment, the step of classifying each river section according to the morphological feature parameters to obtain a plurality of river section types includes: A morphological feature vector including the length, width, tortuosity and slope parameter of the river section is constructed. The morphological feature vector is subjected to cluster analysis, and river sections with similar morphological features are classified into one category to form a plurality of river section types, wherein the river section types include straight type, curved type and valley type. The morphological feature parameters of each type of river section are statistically analyzed to obtain the morphological feature parameter range of each type of river section, and each type of river section is named and defined in combination with the landform features in the remote sensing image.

[0011] In a preferred solution, the step of performing time series analysis on the morphological feature parameters in the sample collection period to identify the morphological feature change trend of each river section comprises: Arranging the morphological feature parameters of each river section according to the collection time sequence to form a time series data set; Analyzing the morphological feature change trend of the river section by time series fitting to determine the change rate of the morphological feature parameters; Taking the absolute value of the change rate of the morphological feature parameters and outputting it as an evaluation index for recording; Comparing the evaluation index with the preset change threshold; When the evaluation index exceeds the change threshold and lasts for N consecutive sampling periods, it is determined that there is a stable morphological change trend in the corresponding river section, and the type of river evolution trend is determined according to the stable morphological change trend, and the corresponding river evolution feature report is output synchronously.

[0012] In a preferred solution, the step of predicting the river morphological feature parameters at the future node based on the morphological feature change trend comprises: Obtaining the demand prediction node and the time interval between the demand prediction node and the current node, and recording it as the prediction time window length; Obtaining the prediction function, and inputting the morphological feature parameters of the current node, the change rate of the morphological feature parameters and the prediction time window length into the prediction function, and recording the output result of the prediction function as the initial prediction parameter; Reasonably checking all initial prediction parameters, eliminating initial prediction parameter combinations that exceed the reasonable range of morphological feature parameters, and retaining initial prediction parameter combinations that simultaneously satisfy the reasonable range of morphological feature parameters; Conducting confidence evaluation on the initial prediction parameter combinations that satisfy the reasonable range of morphological feature parameters, and selecting the initial prediction parameter combination with the highest confidence as the river morphological feature parameters at the demand prediction node.

[0013] The present application also provides a river section morphological feature identification system based on remote sensing, which uses the above-mentioned river section morphological feature identification method based on remote sensing, comprising: A data acquisition module for acquiring multi-source remote sensing data of a target river, wherein the multi-source remote sensing data comprises optical satellite images, synthetic aperture radar images and laser radar point cloud data; A preprocessing module for cooperatively preprocessing the multi-source remote sensing data to obtain a standardized remote sensing data set; A river section division module for extracting a river boundary line and a center line based on the multi-source remote sensing data, and dividing the river into several river sections according to the curvature change characteristics of the river center line; The river section classification module is used for extracting morphological characteristic parameters of each river section, including river section length, width, curvature and river bank slope, and then classifying each river section according to the morphological characteristic parameters to obtain a plurality of river section types, and merging the river sections with consistent types and adjacent to each other to form a continuous river channel morphology module. The river section monitoring module is used for acquiring a repair node of each river section, and constructing a sample collection period between the repair node and a current node, and performing time sequence analysis on the morphological characteristic parameters in the sample collection period to identify the morphological characteristic change trend of each river section, and then predicting the river channel morphological characteristic parameters at a future node based on the morphological characteristic change trend.

[0014] And an electronic device, the electronic device comprises: At least one processor; And a memory connected in communication with the at least one processor; Wherein the memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor to enable the at least one processor to execute the above-mentioned river channel section morphological characteristic identification method based on remote sensing.

[0015] The technical effects obtained by the present application are: The present application realizes accurate identification and dynamic monitoring of river channel morphological characteristics through multi-source remote sensing data fusion analysis, improves the scientificity and efficiency of river channel management, and provides fine data support for river ecological restoration and management through the construction of river section classification and morphology module, so as to support the intelligentization and dataization needs of river channel management decision-making, and through the time sequence analysis of repair node and current node, the effective prediction of river channel morphological evolution trend is realized, and after the prediction result is output, the confidence degree is also checked and evaluated, so as to ensure the reliability and applicability of the prediction result, so that the prediction result can provide corresponding data support for river regulation engineering, and the river channel morphological change trend can be predicted in advance, which also provides a scientific basis for flood control and disaster reduction and water resource management. BRIEF DESCRIPTION OF DRAWINGS

[0016] Figure 1 is a method flowchart of the present application; Figure 2 is a system module schematic diagram of the present application; Figure 3 is an electronic device structure schematic diagram of the present application. DETAILED DESCRIPTION

[0017] In order to make the above-mentioned purposes, features and advantages of the present application more obvious and easy to understand, the specific embodiments of the present application will be described in detail below with reference to the drawings.

[0018] In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present application. However, it will be appreciated that the present application can be practiced in a variety of ways beyond the specific details set forth herein, which can be practiced in other embodiments and applied to other ways within the scope of the present application and therefore the scope of the present application is not limited to the specific embodiments described herein.

[0019] Secondly, the "one embodiment" or "embodiment" referred to herein means that the specific features, structures or characteristics can be included in at least one implementation of the present application. "In a preferred embodiment" appearing in different places in the specification does not mean the same embodiment, nor is it an independent or alternative embodiment.

[0020] Please refer to Figure 1 As shown in the drawings, the present application provides a river section morphological feature recognition method based on remote sensing, comprising: S1, acquiring multi-source remote sensing data of the target river, wherein the multi-source remote sensing data includes optical satellite images, synthetic aperture radar images and laser radar point cloud data; In the step S1, the purpose of river morphology monitoring is mainly to prevent river erosion, siltation and shoreline changes and other problems, and to ensure river flood safety and ecological environment stability. In the present embodiment, first, multi-temporal remote sensing data covering the target river is acquired, including optical satellite images, synthetic aperture radar images and laser radar point cloud data under different seasons and different hydrological conditions, to reflect the morphological change characteristics of the river under various environments. The step of acquiring multi-source remote sensing data of the target river comprises: A plurality of monitoring points are pre-deployed along the river bank, each monitoring point including a collection device for collecting multi-source remote sensing data, the collection device including an optical camera, a synthetic aperture radar antenna and a laser radar scanner; By starting each collection device at a regular time, the optical satellite images, synthetic aperture radar images and laser radar point cloud data of the target river are synchronously acquired, and the initial remote sensing data is summarized; The summarized initial remote sensing data is added with a time stamp and a geographic label to form multi-source remote sensing data with space-time identification; Specifically, when collecting multi-remote sensing data, a plurality of monitoring sites are first pre-deployed along the river bank. These monitoring sites can be uniformly distributed or laid out according to the actual topographic features of the river. The density of monitoring points is appropriately increased in areas where the river morphology changes frequently, such as bends and confluences, to ensure comprehensive monitoring of the target river. Each monitoring site is equipped with an optical camera, a synthetic aperture radar antenna, and a laser radar scanner to capture real-time changes in the river morphology. Then, through a timed collection mechanism, at the preset time point, the collection devices of all monitoring sites are started synchronously to quickly capture optical satellite images, synthetic aperture radar images, and laser radar point cloud data of the target river, and to record them as initial remote sensing data. After the initial remote sensing data is collected, a time stamp and a geographic label are added to it to form a multi-source remote sensing data set with time and space identifiers. The time stamp records the collection time of the data, and the geographic label accurately marks the spatial location of the data, thereby providing a corresponding data basis for subsequent analysis and processing.

[0021] S2, cooperatively pre-processing the multi-source remote sensing data to obtain a standardized remote sensing data set; In the step S2, after the multi-source remote sensing data is output, each source remote sensing data is cooperatively pre-processed to eliminate errors and noises caused by sensor differences, atmospheric interference, and terrain undulations during data acquisition, thereby ensuring the consistency and comparability of the data. The step of cooperatively pre-processing the multi-source remote sensing data includes: unifying the spatial reference of the optical satellite images, the synthetic aperture radar images, and the laser radar point cloud data, and registering the optical satellite images, the synthetic aperture radar images, and the laser radar point cloud data to a unified coordinate system; performing multi-temporal image superposition and band reconstruction on the optical satellite images to recover the obscured areas; performing denoising and geometric correction on the synthetic aperture radar images; performing filtering and elevation normalization on the laser radar point cloud data; fusing the pre-processed optical satellite images, synthetic aperture radar images, and laser radar point cloud data to integrate them into a unified information layer, thereby forming a standardized remote sensing data set with a unified spatial reference; Specifically, in the collaborative preprocessing of multi-source remote sensing data, first, the spatial reference of optical satellite images, synthetic aperture radar images and laser radar point cloud data is unified, which is registered to the same coordinate system to ensure the spatial consistency between the data and avoid misinterpretation or incorrect conclusions caused by direct comparison or analysis. Then, multi-temporal image overlay and band reconstruction are performed on the optical satellite images to recover the areas obscured by clouds, shadows, etc. Multi-temporal image overlay can use image data at different time points to identify obscured areas that are visible at other time points through comparison and analysis, thereby recovering the information of the corresponding areas. Band reconstruction enhances the information extraction capability of specific features such as water bodies and vegetation by adjusting the band combination of the images. For synthetic aperture radar images, denoising and geometric correction are mainly performed. Denoising is to eliminate speckle noise in the image to improve image clarity. Geometric correction is to correct geometric distortion in the image to ensure that the shape, size and direction of features in the image match the actual situation. Laser radar point cloud data needs to be filtered and elevation normalized. Filtering is to remove noise points in the point cloud data such as vegetation and buildings to obtain more accurate ground elevation information. Elevation normalization is to unify the elevation values of the point cloud data to the same reference surface to facilitate subsequent data analysis and processing. Then, data fusion is performed on the optical satellite images, synthetic aperture radar images and laser radar point cloud data to integrate them into a unified information layer, thereby integrating remote sensing data from different sources into a standardized remote sensing data set with a unified spatial reference.

[0022] S3, extracting a river boundary line and a center line based on the multi-source remote sensing data, and dividing the river into a plurality of river sections according to the curvature variation characteristics of the river center line; In the step S3, after the multi-source remote sensing data is output and preprocessed, the river boundary line and the center line can be extracted based on the multi-source remote sensing data to provide corresponding pre-data support for river section division. The step of extracting the river boundary line and the center line based on the multi-source remote sensing data and dividing the river into a plurality of river sections according to the curvature variation characteristics of the river center line includes: calculating a river water mask based on the spectral characteristics of the optical satellite images, and correcting the boundary of the water mask in combination with the microwave reflection characteristics of the synthetic aperture radar images; extracting river terrain elevation information using the laser radar point cloud data, and determining a river bank line according to the slope variation; extracting a river boundary line according to the fusion of the river water mask and the river bank line information, and extracting a river center line in combination with the river boundary line and the terrain elevation data; calculating the curvature of the river center line, and marking the curvature peak points; The feature points exceeding two times of the standard deviation of the overall curvature fluctuation range and having a distance between adjacent curvature peak points greater than the average width of the river channel are extracted as the river section division nodes; The river channel is divided into several independent river sections according to the river section division nodes, and the boundary range of each independent river section is extracted based on the center line information of the independent river section; Specifically, in the process of dividing the river section, first, the water body is extracted from the optical satellite image based on the standardized remote sensing data set by using image processing technology (the image processing technology can use common image segmentation algorithms such as deep learning models such as U-Net or Mask R-CNN, etc.), to obtain a preliminary water body mask. Due to the influence of factors such as light and shadow, the boundary of the water body mask may not be accurate enough, so the microwave reflection characteristics of the synthetic aperture radar image are combined to correct the boundary of the water body mask to improve the extraction accuracy of the river boundary line. Then, the elevation information in the laser radar point cloud data is used to determine the elevation range of the river bank line through terrain analysis. The river bank line is the dividing line between the river channel and the land. Then, the water body mask and the river bank line information are fused to extract a more accurate river boundary line through intersection or fusion processing. At the same time, the river center line can be extracted by combining the river boundary line and the terrain elevation data. The river center line is the center axis of the river shape, and its position and shape reflect the overall trend and bending degree of the river. Then, the curvature of the river center line is calculated to identify the bending feature points of the river, i.e. the curvature peak points. Through analysis of the curvature peak points, the feature points exceeding two times of the standard deviation of the overall curvature fluctuation range and having a distance between adjacent curvature peak points greater than the average width of the river channel are extracted as the river section division nodes. Based on this, the river channel is divided into several independent river sections. Finally, based on the center line information of each independent river section, the boundary range of the corresponding river section can be extracted (specifically, the center line can be smoothed by a spatial interpolation method, and the curvature is calculated based on the smoothed center line to eliminate the influence of local noise on the identification of curvature peak points), which provides basic data support for subsequent river section classification, monitoring and prediction.

[0023] S4, the morphological feature parameters of each river section are extracted, including the length, width, bending degree and bank slope of the river section, and then each river section is classified according to the morphological feature parameters to obtain a plurality of river section types, and the river sections with the same type and adjacent to each other are merged to form a continuous river shape module; In the step S4, after the river section division is completed, the morphological characteristic parameters of each river section are extracted, including the river section length, width, curvature and bank slope, and then each river section is classified according to the preset classification standard, such as the length range, curvature level and bank slope category. The purpose of classification is to classify the river sections with similar morphological characteristics into one category, so as to facilitate subsequent analysis and management. After the river section classification is completed, the river sections of the same type and adjacent to each other are merged to form continuous river channel morphological modules, so as to realize the systematic cognition of the overall river channel structure. The step of extracting the morphological characteristic parameters of each river section includes: The river section length is calculated by accumulating the distance between adjacent coordinate points based on the coordinate sequence of the center line of each river section. The normal profile line is generated along the center line of the river section, and the intersection distance between the river boundary line and the normal profile line is used to determine the river section width. The straight line distance between the starting point and the ending point of the center line of the river section is obtained, and the curvature of the river section is calculated in combination with the actual length of the river section. The value of the curvature is the ratio of the actual length of the river section to the straight line distance between the starting point and the ending point of the center line of the river section. The elevation points of the slope surface on both banks are extracted along the river boundary line, the slope angle of the slope surface is fitted by linear regression, and the average value of the slope on both banks is taken as the slope parameter of the river section. The river section length, width, curvature and river section slope parameter are recorded as morphological characteristic parameters. Specifically, when the morphological characteristic parameters of the river section are extracted, the length of the river section is first calculated by accumulating the distance between adjacent coordinate points based on the coordinate sequence of the center line of each river section. Then the normal profile line is generated along the center line of the river section, and the intersection distance between the normal profile line and the river boundary line is used to determine the river section width, so as to capture the width variation of the river section at different positions and understand the morphological characteristics of the river channel. Then the straight line distance between the starting point and the ending point of the center line of the river section is obtained, and the curvature of the river section is calculated in combination with the actual length of the river section. The quantitative value of the curvature is the ratio of the actual length of the river section to the straight line distance between the starting point and the ending point of the center line of the river section. The larger the ratio, the higher the curvature of the river channel. At the same time, the elevation points of the slope surface on both banks are extracted along the river boundary line, and the slope angle of the slope surface is fitted by linear regression. Finally, the average value of the slope on both banks is taken as the slope parameter of the river section. Finally, the river section length, width, curvature and river section slope parameter are recorded as morphological characteristic parameters.

[0024] Secondly, the step of classifying each river section according to the morphological characteristic parameters to obtain a plurality of river section types includes: A morphological characteristic vector including the river section length, width, curvature and slope parameter is constructed. The morphological feature vectors are subjected to cluster analysis, and river sections with similar morphological features are classified into one category to form multiple river section types, wherein the river section types include straight type, curved type and canyon type; The morphological feature parameters of each type of river section are subjected to statistical analysis to obtain the morphological feature parameter ranges of each type of river section, and each type of river section is named and defined in combination with the landform features in the remote sensing image. In the above, when classifying the river sections, the morphological feature parameters of the river sections, including the river section length, width, curvature and slope parameters, are first integrated into morphological feature vectors, which can reflect the geometric and topographic features of the river sections and provide data basis for subsequent classification. Then, a cluster analysis algorithm such as K-means or DBSCAN is used to cluster the morphological feature vectors. The cluster analysis can automatically classify river sections with similar features into one category without pre-setting classification criteria, so that the classification result is more objective. Through cluster analysis, multiple river section types such as straight type, curved type and canyon type can be obtained. Straight type river sections usually have straight river courses with small changes in width and slope. Curved type river sections have obvious curved morphology with high curvature. Canyon type river sections are located in narrow valleys with steep riverbank slopes. In order to more accurately describe and define each type of river section, statistical analysis of the morphological feature parameters of each type of river section is performed to obtain the morphological feature parameter ranges of each type of river section. At the same time, in combination with the landform features in the remote sensing image, each type of river section is named and defined, such as naming the river sections with straight river courses and uniform width as “straight type river sections”, and naming the river sections with high curvature and complex morphology as “complex curved type river sections”, etc. This helps to have a clear understanding and definition of the river morphology features.

[0025] S5, obtaining the repair nodes of each river section, and constructing a sample collection period between the repair node and the current node, and performing time series analysis on the morphological feature parameters in the sample collection period to identify the morphological feature change trend of each river section, and then predicting the river morphology feature parameters at the future node based on the morphological feature change trend.

[0026] In the step S5, after the river section division is completed, the dynamic monitoring of the morphological characteristic parameters is performed, specifically, the node of the last repair operation is selected as the repair node, and the time period between the repair node and the current node is defined as the sample collection period, so as to perform the time series analysis on the change of the river section morphological characteristic parameters in the time period, through the time series analysis, the morphological characteristic change trend of the river section can be recognized, for example, the change rate of the river width, the evolution trend of the bending degree or the adjustment direction of the slope, etc., based on the recognized change trend, the morphological characteristic parameters of the river section at a specific time node in the future are determined, thereby providing the corresponding basis for the subsequent river management and ecological restoration, wherein, the step of performing the time series analysis on the morphological characteristic parameters in the sample collection period and recognizing the morphological characteristic change trend of each river section, comprises: arranging the morphological characteristic parameters of each river section according to the collection time sequence to form a time series data set; analyzing the morphological characteristic change trend of the river section through the time series fitting mode to determine the change rate of the morphological characteristic parameters; performing the absolute value processing on the change rate of the morphological characteristic parameters, and outputting the evaluation index for recording; comparing the evaluation index with the preset change threshold; when the evaluation index exceeds the change threshold and continuously lasts for N sampling periods, it is determined that there is a stable morphological change trend in the corresponding river section, and according to the stable morphological change trend, the type of the river section evolution trend is determined, and the corresponding river evolution characteristic report is outputted synchronously; Specifically, in determining the morphological feature change trend of the river section, first, the morphological feature parameters of each river section, such as length, width, curvature and slope parameters, are arranged according to the time sequence collected to form a time series data set, which can reflect the change of the river section morphological feature parameters with time. Then, the time series fitting method, such as ARIMA model or LSTM neural network, is used to analyze the morphological feature change trend of the river section, and the possible change of the future river section morphological feature parameters is predicted by fitting the historical data, so as to determine the change rate of the morphological feature parameters (i.e. the value of the morphological feature parameters change per unit time), and thus determine the evolution speed and direction of the river section morphological feature. Then, the absolute value of the change rate of the morphological feature parameters is processed to eliminate the influence of the sign on the change degree, and the processed change rate is recorded as an evaluation index for subsequent comparison and analysis. After that, the evaluation index is compared with the preset change threshold, which is set according to the natural evolution law of the river channel, historical repair experience and other factors, to judge whether the change of the river section morphological feature parameters is significant. When the evaluation index exceeds the change threshold and remains in the exceeding state for N consecutive sampling periods (N is the preset sampling period number, generally between 3-5), it is determined that there is a stable morphological change trend in the corresponding river section, which indicates that the morphological feature of the corresponding river section is changing at a certain speed and direction. Finally, the type of river evolution trend is determined according to the stable morphological change trend, such as river widening trend, curvature increasing trend or slope slowing down trend, and the corresponding river evolution feature report is output synchronously.

[0027] Secondly, the step of predicting the river channel morphological feature parameters at the future node based on the morphological feature change trend includes: Obtain the demand prediction node and the time interval between the demand prediction node and the current node, and record it as the prediction time window length; Obtain the prediction function, and input the morphological feature parameters of the current node, the change rate of the morphological feature parameters and the prediction time window length into the prediction function, and record the output result of the prediction function as the initial prediction parameter; Reasonably check all initial prediction parameters, eliminate the initial prediction parameter combination that exceeds the reasonable range of morphological feature parameters, and retain the initial prediction parameter combination that meets the reasonable range of morphological feature parameters at the same time; Conduct confidence evaluation on the initial prediction parameter combination that meets the reasonable range of morphological feature parameters, and select the initial prediction parameter combination with the highest confidence as the river channel morphological feature parameters at the demand prediction node.

[0028] Specifically, after determining the morphological change trend of the river section, in order to predict the river morphological feature parameters at the future node based on the morphological feature change trend, first of all, the demand prediction node is determined, which represents the time point at which the future river morphological features need to be predicted. At the same time, the time interval between the demand prediction node and the current node is calculated and recorded as the prediction time window length. Then, a preset prediction function is introduced, which can estimate the river morphological feature parameters at the future node based on historical data (i.e. the morphological feature parameters of the current node, the change rate of the morphological feature parameters) and the prediction time window length. After inputting the relevant data into the prediction function, the output result obtained is the initial prediction parameter. The expression of the prediction function is: initial prediction parameter = current morphological feature parameter + (morphological feature parameter change rate x prediction time window length). The initial prediction parameter is a preliminary estimate of the future river morphological feature parameters. However, due to possible errors or uncertainties in the prediction process, all initial prediction parameters need to be reasonably checked. The standard for checking is the reasonable range of morphological feature parameters, which is based on prior knowledge of river morphology and statistical analysis of historical data. By eliminating initial prediction parameter combinations that exceed the reasonable range, prediction results that are more likely to be close to the actual situation can be retained. Further, in order to select the optimal solution from the retained initial prediction parameter combinations, confidence evaluation is performed. Confidence evaluation considers the reliability and accuracy of the prediction results. By comparing the confidence levels of different initial prediction parameter combinations, the combination with the highest confidence level is selected as the river morphological feature parameters at the demand prediction node. The confidence level evaluation method includes comprehensive consideration of factors such as the historical performance of the prediction model, the stability of parameter changes, and the influence of the prediction time window length. Through quantitative analysis (such as linear regression analysis or weighted average method) of the confidence levels of multiple prediction parameter combinations, the initial prediction parameter combination that is most likely to reflect the future actual river morphology can be effectively identified. In this way, not only the reliability of the prediction results is improved, but also the bias risk caused by a single prediction method is reduced, making the prediction results of the river section morphological feature parameters more accurate, so as to provide a scientific basis for subsequent river management, flood control planning, and ecological protection measures.

[0029] Please refer to Figure 2 A river section morphological feature recognition system based on remote sensing using the above-mentioned river section morphological feature recognition method based on remote sensing, comprising: A data acquisition module for acquiring multi-source remote sensing data of a target river, wherein the multi-source remote sensing data includes optical satellite images, synthetic aperture radar images, and laser radar point cloud data; A preprocessing module for cooperatively preprocessing the multi-source remote sensing data to obtain a standardized remote sensing data set; The river section division module is configured to extract a river boundary line and a center line based on multi-source remote sensing data, and divide the river into a plurality of river sections according to a curvature variation characteristic of the river center line. The river section classification module is configured to extract morphological characteristic parameters of each river section, including a river section length, a width, a curvature, and a bank slope, and classify each river section according to the morphological characteristic parameters to obtain a plurality of river section types, and merge river sections with consistent river section types and adjacent river sections to form a continuous river morphological module. The river section monitoring module is configured to obtain a repair node of each river section, and construct a sample collection time period between the repair node and a current node, and perform time series analysis on morphological characteristic parameters in the sample collection time period to identify a morphological characteristic variation trend of each river section, and predict a river morphological characteristic parameter at a future node based on the morphological characteristic variation trend.

[0030] The execution process of the identification system corresponds to the implementation steps of the above-mentioned remote sensing-based river section morphological characteristic identification method, and repetitive repetition is not performed here.

[0031] Please refer to Figure 3 An electronic device includes: at least one processor; and a memory connected in communication with the at least one processor; wherein the memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor to enable the at least one processor to execute the above-mentioned remote sensing-based river section morphological characteristic identification method.

[0032] The processor of the above-mentioned electronic device can be a central processing unit (CPU), a graphics processing unit (GPU), or other computing cores suitable for data processing, and the memory can be a random access memory (RAM), a read-only memory (ROM), or a solid state drive (SSD) storage medium, in addition, the electronic device can also include an arithmetic logic unit (ALU), a controller, and an input / output interface for supporting efficient processing of the processor and external communication interaction.

[0033] It should be noted that in this paper, the terms "include", "contain" or any other variant thereof are intended to cover non-exclusive inclusion, so that the process, device, article or method including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such process, device, article or method. Without more limitations, the element defined by the statement "includes a" does not exclude the presence of another identical element in the process, device, article or method including the element.

[0034] The above merely describes the preferred embodiments of the present application, and it should be pointed out that those skilled in the art can make several improvements and refinements without departing from the principles of the present application, and these improvements and refinements should also be considered as the protection scope of the present application. The structures, devices and operation methods not specifically described and explained in the present application are implemented according to the conventional means in the art, unless specifically described and limited.

Claims

1. A remote sensing-based river reach morphological feature identification method, characterized in that: The method comprises the steps of: acquiring multi-source remote sensing data of a target river channel, wherein the multi-source remote sensing data comprises optical satellite images, synthetic aperture radar images and laser radar point cloud data; cooperatively preprocessing the multi-source remote sensing data to obtain a standardized remote sensing data set; extracting a river channel boundary line and a center line based on the multi-source remote sensing data, and dividing the river channel into a plurality of river sections according to the curvature variation characteristics of the river channel center line; extracting morphological characteristic parameters of each river section, including river section length, width, curvature and river bank slope, and then classifying each river section according to the morphological characteristic parameters to obtain a plurality of river section types, and merging river sections with consistent river section types and adjacent river sections to form continuous river channel morphological modules; acquiring a repair node of each river section, and constructing a sample collection time period between the repair node and a current node, and performing time series analysis on the morphological characteristic parameters in the sample collection time period to identify the morphological characteristic variation trend of each river section, and then predicting the river channel morphological characteristic parameters at a future node based on the morphological characteristic variation trend.

2. The method according to claim 1, characterized in that: The step of acquiring multi-source remote sensing data of a target river channel comprises: pre-deploying a plurality of monitoring points along the river bank, each monitoring point comprising a collection device for collecting multi-source remote sensing data, the collection device comprising an optical camera, a synthetic aperture radar antenna and a laser radar scanner; synchronously acquiring optical satellite images, synthetic aperture radar images and laser radar point cloud data of the target river channel by starting each collection device at a fixed time, and collecting initial remote sensing data; adding time stamps and geographical labels to the collected initial remote sensing data to form multi-source remote sensing data with space-time identification.

3. The method according to claim 1, characterized in that: The step of cooperatively preprocessing the multi-source remote sensing data comprises: unifying the spatial reference of the optical satellite images, the synthetic aperture radar images and the laser radar point cloud data, and registering the optical satellite images, the synthetic aperture radar images and the laser radar point cloud data to a unified coordinate system; performing multi-temporal image superposition and band reconstruction on the optical satellite images to recover the hidden areas; performing denoising and geometric correction on the synthetic aperture radar images; performing filtering and elevation normalization on the laser radar point cloud data; data fusion of the preprocessed optical satellite images, synthetic aperture radar images and laser radar point cloud data to integrate them into a unified information layer to form a standardized remote sensing data set with a unified spatial reference.

4. The method according to claim 3, characterized in that: The step of extracting a river channel boundary line and a center line based on multi-source remote sensing data, and dividing the river channel into a plurality of river sections according to the curvature variation characteristics of the river channel center line comprises: calculating a river water mask based on the spectral characteristics of the optical satellite images, and combining the microwave reflection characteristics of the synthetic aperture radar images to correct the boundaries of the water mask; extracting river terrain elevation information from the laser radar point cloud data, and determining the river bank line according to the slope variation; extracting the river channel boundary line according to the fusion of the river water mask and the river bank line information, and extracting the river channel center line in combination with the river channel boundary line and the terrain elevation data; calculating the curvature of the river channel center line, and marking the curvature peak points; extracting feature points with a curvature peak point that exceeds twice the standard deviation of the overall curvature fluctuation range and a distance between adjacent curvature peak points greater than the average width of the river channel as river section division nodes; The river course is divided into several independent river sections according to the river section division node, and the boundary range of each independent river section is extracted based on the center line information of the independent river section.

5. The method according to claim 1, characterized in that: The step of extracting the morphological characteristic parameters of each river section comprises: calculating the length of the river section by accumulating the distance between adjacent coordinate points based on the coordinate sequence of the center line of the river section; generating a normal profile line along the center line of the river section, and determining the width of the river section according to the distance between the intersection points of the river course boundary line and the normal profile line; obtaining the straight-line distance between the starting point and the ending point of the center line of the river section, and calculating the tortuosity of the river section in combination with the actual length of the river section, wherein the value of the tortuosity is the ratio of the actual length of the river section to the straight-line distance between the starting point and the ending point of the center line of the river section; extracting the elevation points of the slope surface on both banks along the river course boundary line, fitting the slope angle by linear regression, and taking the average of the slope angles on both banks as the slope parameter of the river section; recording the length, width, tortuosity and slope parameter of the river section as the morphological characteristic parameters.

6. The method according to claim 1, characterized in that: The step of classifying each river section according to the morphological characteristic parameters to obtain a plurality of river section types comprises: constructing a morphological characteristic vector comprising the length, width, tortuosity and slope parameter of the river section; performing cluster analysis on the morphological characteristic vector, grouping the river sections with similar morphological characteristics into one category, and forming a plurality of river section types, wherein the river section types include straight-line type, curved type and canyon type; performing statistical analysis on the morphological characteristic parameters of each type of river section, obtaining the range of morphological characteristic parameters of each type of river section, and naming and defining each type of river section in combination with the landform features in the remote sensing image.

7. The method according to claim 1, characterized in that: The step of performing time series analysis on the morphological characteristic parameters in the sample collection period to identify the morphological characteristic change trend of each river section comprises: arranging the morphological characteristic parameters of each river section according to the collection time sequence to form a time series data set; analyzing the morphological characteristic change trend of the river section by time series fitting, and determining the change rate of the morphological characteristic parameter; taking the absolute value of the change rate of the morphological characteristic parameter, and outputting it as an evaluation index for recording; comparing the evaluation index with the preset change threshold; when the evaluation index exceeds the change threshold and lasts for N consecutive sampling periods, it is determined that there is a stable morphological change trend in the corresponding river section, and the type of the river section evolution trend is determined according to the stable morphological change trend, and the corresponding river evolution characteristic report is outputted synchronously. 8.The method of claim 1, wherein the method further comprises: determining a river section shape feature of the river section based on the river section shape feature recognition result. The step of predicting the morphological characteristic parameters of the river course at the future node based on the morphological characteristic change trend comprises: obtaining the demand prediction node, the time interval between the demand prediction node and the current node, and recording the prediction time window length; obtaining the prediction function, inputting the morphological characteristic parameters of the current node, the change rate of the morphological characteristic parameters and the prediction time window length into the prediction function, and recording the output result of the prediction function as the initial prediction parameter; performing reasonableness verification on all the initial prediction parameters, eliminating the initial prediction parameter combinations that exceed the reasonable range of the morphological characteristic parameters, and retaining the initial prediction parameter combinations that simultaneously satisfy the reasonable range of the morphological characteristic parameters; The initial prediction parameter combination satisfying the reasonable range of the morphological feature parameter is subjected to confidence evaluation, and the initial prediction parameter combination with the highest confidence is selected as the river morphological feature parameter under the demand prediction node.

9. A remote sensing-based river reach morphological feature identification system, characterized by: The method comprises the following steps: A data acquisition module is configured to acquire multi-source remote sensing data of a target river, wherein the multi-source remote sensing data comprises optical satellite images, synthetic aperture radar images and laser radar point cloud data. A preprocessing module is configured to perform collaborative preprocessing on the multi-source remote sensing data to obtain a standardized remote sensing data set. A river section division module is configured to extract a river boundary line and a center line based on the multi-source remote sensing data, and divide the river into a plurality of river sections according to the curvature variation characteristics of the river center line. A river section classification module is configured to extract morphological feature parameters of each river section, including river section length, width, curvature and bank slope, and then classify each river section according to the morphological feature parameters to obtain a plurality of river section types, and merge river sections with consistent types and adjacent to each other to form continuous river morphological modules. A river section monitoring module is configured to acquire a repair node of each river section, and construct a sample collection period between the repair node and a current node, and perform time series analysis on the morphological feature parameters in the sample collection period to identify the morphological feature change trend of each river section, and then predict the river morphological feature parameters at a future node based on the morphological feature change trend.

10. An electronic device, comprising: The electronic device comprises: at least one processor; and a memory connected in communication with the at least one processor; wherein the memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor to enable the at least one processor to execute the remote sensing-based river section morphological feature recognition method of any one of claims 1 to 8.

Citation Information

Patent Citations

  • River and lake demarcation identification monitoring method based on remote sensing data

    CN117423010A

  • River reach form recognition method and system, storage medium and electronic equipment

    CN119649054A

  • Dynamic change monitoring method and system based on multi-temporal river channel image

    CN120356105A

  • Monitoring method and device base on aerospace multi-source remote sensing data

    WO2018098738A1

Cited By

  • Cognitive disorder risk intelligent matching intervention system based on multi-modal data

    CN121859268A