Method for evaluating stability of wandering river regime based on trajectory similarity algorithm

Abstract

The application discloses a meandering river regime stability evaluation method based on a trajectory similarity algorithm, and aims at the defects that the existing river regime stability evaluation method is single in indexes and difficult to comprehensively reflect the overall stability degree of the river regime and the local abnormal river regime, the normalized Frechet distance is selected to represent the overall shape similarity degree of the river regime, the bidirectional Hausdorff distance is selected to represent the local maximum deviation degree of the river regime, and a two-dimensional river regime stability evaluation method of 'overall similarity-local difference' is constructed based on the trajectory similarity algorithm, so that the standardization pretreatment of the river mainstream line vector data can be carried out, the long-term overall shape evolution law of the river regime can be analyzed, the deviation degree of the actual river regime and the planning regulation line can be analyzed, and quantitative technical support can be provided for the layout optimization of the meandering river regulation project and the risk identification of the local abnormal river regime.

Description

Technical Field

[0001] This invention relates to the field of river regulation technology in water conservancy projects, and is particularly applicable to a method for evaluating the stability of wandering rivers based on trajectory similarity algorithms. Background Technology

[0002] Wandering rivers are characterized by wide, shallow, and scattered riverbeds, numerous mid-channel sandbars, violent lateral swings of the main stream, dramatic scouring and sedimentation, and extremely poor river stability. They are mostly developed in plains with abundant sediment supply, large gradients, and weak bank erosion resistance, and are distributed across all continents of the world.

[0003] Wandering rivers are characterized by frequent main channel shifts, difficulty in controlling their course, and a high risk of hazard. Many scholars have conducted research on the evolution trends and management strategies of wandering rivers. When analyzing the evolution of wandering river courses, it is crucial to select appropriate indicators to reflect the overall morphological evolution and stability of the river course. Traditional indicators such as tortuosity, lateral sway intensity, radius of curvature, and wandering index only reflect single characteristics of the river course and cannot comprehensively depict the overall topological morphology and local abnormal river course development. Currently, there is no standardized method for evaluating river course stability that considers both the overall morphological evolution and the identification of local abnormal river course, thus failing to meet the practical engineering needs of wandering river regulation and flood control. Summary of the Invention

[0004] The purpose of this invention is to provide a method for evaluating the stability of wandering rivers based on trajectory similarity algorithm. This method addresses the problem of evaluating the stability of wandering rivers where the overall topology and local abnormal river development cannot be fully characterized. It provides a scientific basis for analyzing the evolution of wandering rivers, identifying risk areas of abnormal river conditions, optimizing the layout of river regulation projects, and revising planning guidelines.

[0005] To achieve the above objectives, the method for evaluating the river morphology stability of wandering rivers based on trajectory similarity algorithm described in this invention includes the following steps: S1, River vector data preprocessing: Obtain river vector data of different periods of the wandering river section to be evaluated, remove outliers, unify coordinate system and sequence direction, and resample at equal intervals to obtain the discrete point sequence coordinates [X,Y] arranged unidirectionally from upstream to downstream of the river at different periods, ensuring that the start and end points of the sequence correspond to the starting and ending sections of the wandering river section to be evaluated, respectively. S2, Discrete Fraser distance calculation: Based on any two pre-processed sets of unidirectionally ordered discrete point sequences, the discrete Fraser distance is calculated using a dynamic programming algorithm, and then normalized by combining it with the actual length of the benchmark river channel to obtain the normalized Fraser distance, which is used to characterize the evolution process and laws of the overall topological morphology of the river. S3, Bidirectional Hausdorff distance calculation: Based on any two preprocessed sets of unidirectionally ordered discrete point sequences, calculate the bidirectional directed Hausdorff distance, and take the maximum value of the two directed distances as the final Hausdorff distance. S4, Two-Dimensional Collaborative Joint Evaluation: Based on the normalized Fraser distance and the two-way Hausdorff distance, a two-dimensional matrix joint evaluation system is constructed. Combined with the pre-set river stability evaluation grading standards, the final river status evaluation results are output.

[0006] Furthermore, in step S1, the river vector line data is the main channel vector data or the planning and guiding line vector data of the river regulation project. The data sources include river morphology survey reports, remote sensing impact interpretation, large cross-section observation data, and approved river regulation line data. The coordinate system is the Gauss-Kruger 3° plane projection coordinate system to eliminate the influence of geographic coordinate projection on distance calculation.

[0007] Furthermore, in step S1, the outlier removal uses a sliding window. The criteria are as follows: using 5 consecutive sampling points as a sliding window, calculate the mean and standard deviation of the horizontal offset of the points within the window, and remove outliers that deviate from the mean by more than a preset threshold.

[0008] Furthermore, in step S2, The formula for calculating the normalized Fréchet distance is as follows: In the formula, The actual length of the river channel vector line during the baseline period is normalized to a dimensionless value of 0-1. It is the discrete Friesian distance between the preprocessed, unidirectionally ordered discrete point sequence P used as the benchmark and the preprocessed, unidirectionally ordered discrete point sequence Q used for comparison.

[0009] Furthermore, in step S3, the formula for calculating the bidirectional directed Hausdorff distance is: Where P is the preprocessed baseline period unidirectional ordered discrete point sequence, and Q is the preprocessed comparison period unidirectional ordered discrete point sequence. The directed Hausdorff distance from P to Q is calculated using the following formula: In the formula, The Euclidean distance between the two points; Let be the directed Hausdorff distance from Q to P. The formula for calculation is: In the formula, This is the Euclidean distance between two points.

[0010] Furthermore, in step S4, the river state classification standard of the dual-dimensional matrix joint evaluation system is specifically as follows: Furthermore, the two sets of unidirectionally ordered discrete point sequences are measured mainstream river data from different periods in the same river segment, used for analyzing the interannual evolution of river morphology; or the two sets of unidirectionally ordered discrete point sequences are mainstream river vector data and planned traverse line data in the same river segment, used for analyzing the degree of deviation between river morphology and planned traverse line.

[0011] The advantages of this invention are: By combining the Fraser distance and the two-way Hausdorf distance, and comparing the computational logic and applicable scope of the two algorithms, a river stability evaluation index system for wandering rivers was formed. This system uses the normalized Fraser distance as the core to accurately characterize the evolution of the overall topological morphology of the river, and uses the two-way Hausdorf distance as an auxiliary to effectively identify the flood control risks of local extreme swings. This solves the core technical problem that existing single indicators cannot simultaneously take into account both overall morphological stability and local risk identification.

[0012] Through a complete technical process of "data preprocessing - dual-index collaborative calculation - hierarchical evaluation", the quantitative calculation and analysis of river morphology has been realized. It can be used to analyze the overall morphological evolution of historical river morphology in different periods, as well as to analyze the degree of deviation between the measured river morphology and the planned guiding line. The application scenarios are quite extensive. Attached Figure Description

[0013] Figure 1 This invention relates to a method for evaluating the stability of wandering river channels based on a trajectory similarity algorithm.

[0014] Figure 2 This is a comparison of the measured main current lines of the wandering river section in the lower reaches of the Yellow River in 2020 and 2021, as described in Example 1.

[0015] Figure 3 This is a comparison chart of the measured main channel line and the planned guiding line of the wandering river section in the lower reaches of the Yellow River in 2020, as described in Example 2. Detailed Implementation

[0016] The technical solutions in the embodiments of the present invention will be clearly and completely described below. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0017] Example 1 The river section to be evaluated in this embodiment is a section of the lower Yellow River. This section is a typical meandering river section, with a total length of 299 km, a bank-to-bank distance of 4.1 to 20.0 km, and a channel gradient between 0.172 and 0.256‰. Based on the pre-flood season vector data of the main channel in 2020 and 2021, this embodiment calculates and analyzes the normalized Fraser distance and the two-way Hausdorff distance of this river section to evaluate the overall changes in the river morphology and the location of the maximum local swing in this meandering river section.

[0018] like Figure 1 As shown, the specific implementation steps of the wandering river channel stability evaluation method based on trajectory similarity algorithm described in this invention are as follows: S1, River vector data preprocessing, to obtain the main channel vector data of the wandering river section to be evaluated at different periods.

[0019] (1) Data acquisition: The main channel lines in the pre-flood season river regime survey reports of the lower Yellow River in 2020 and 2021 were vectorized using georeferencing. For example... Figure 2 The image shows a comparison of the measured main channel of the Yellow River in the lower reaches before the flood season in 2020 and 2021.

[0020] (2) Outlier removal: The sliding window 3σ criterion is adopted, with 5 consecutive sampling points as the window, to remove outliers such as river islands and shoals that deviate from the mean by more than the preset threshold, such as 3 times the standard deviation; (3) Coordinate system one: The two main lines are projected and transformed. The projection coordinate system is the Gauss-Kruger 3° zone plane projection coordinate system with a central longitude of 114°. (4) Uniform sequence direction: The main lines of both phases are adjusted to a unidirectional orderly arrangement from upstream to downstream to ensure that the starting and ending sections of the river sections corresponding to the starting and ending points of the sequence are completely consistent; (5) Equal-interval resampling: For this wandering river section, the two main current lines were resampled at equal intervals of 100m, converting the main current line vector data into discrete points. Specifically, the 2020 main current line sequence P was discretized to obtain 3639 points. The X and Y coordinates of different discrete points were calculated, and the point set P = {p1, p2, ..., p...} m}, where p i =(x i ,y i Let q1 be a plane coordinate point; in 2021, the mainstream line sequence Q was discretized to obtain 3542 points, and the X and Y coordinates of different discrete points were calculated, resulting in the point set Q = {q1, q2, q3}. q2 ,...,q m}, where q i =(xi ,y i Let P be a unidirectional ordered discrete point sequence in the baseline period, and Q be a unidirectional ordered discrete point sequence in the comparison period.

[0021] S2, Discrete Fraser Distance Calculation: Based on two sets of unidirectional ordered discrete point sequences after preprocessing the mainstream vector data of the river channel at different periods, a 3649×3542 Euclidean distance matrix is ​​constructed. Following a dynamic programming algorithm, the original Fraser distance d is calculated layer by layer according to the river direction to obtain the denominator. F =2483m.

[0022] Where P is the preprocessed baseline unidirectional ordered discrete point sequence, and Q is the preprocessed comparison period unidirectional ordered discrete point sequence; and are monotonically non-decreasing mapping functions that satisfy the sequence order constraints of P and Q, respectively; refers to the maximum distance between corresponding points on two unidirectional ordered discrete point sequences during the entire unidirectional ordered movement of the parameter t of the monotonically non-decreasing mapping function from 0 to 1. This maximum value is the maximum value under one type of movement; refers to the minimum value among all legal movements, that is, the optimal movement that minimizes the velocity of the moving points on the two trajectories.

[0023] Combined with the total length L of the river section P =299000m, the normalized Fréchet distance d is calculated. F-norm =0.008, the result corresponds to the same overall state of the river channel.

[0024] S3, Hausdorff distance calculation: Iterate through all 3649 points on the main trend line in 2020, and match them one by one with the nearest points on the main trend line in 2021 to obtain the directed Hausdorff distance from P to Q. That is, the maximum value of all shortest distances from P to Q is d. H (P,Q) = 2483m; By iterating backwards through all 3542 points in 2021 and matching them with the river channel sequence in 2020, the directed Hausdorff distance from Q to P, i.e., the reverse one-way distance d, is obtained. H（Q,P） =1877m; taking the maximum of the two, the final bidirectional Hausdorff distance d in the river segment is . H =2483m, which indicates that the maximum lateral swing distance of the main channel in this river section in 2020-2021 was 2483m, and the result corresponds to obvious local swing of the river channel.

[0025] S4, Two-Dimensional Collaborative Joint Evaluation: Based on normalized Friesian distance and bidirectional Hausdorff distance, a two-dimensional matrix-based joint evaluation system of "overall similarity-local dissimilarity" is constructed. The specific river morphology classification standard of the two-dimensional matrix-based joint evaluation system is as follows: According to the river state grading standard of the dual-dimensional matrix joint evaluation system, the normalized Fraser distance of 0.008 < 0.01 indicates that the overall river state in 2020 and 2021 was similar. The two-way Hausdorf distance of 2483m is greater than 1000m, corresponding to significant local river channel undulation. Comparison with remote sensing images reveals that this undulation location was near a certain control project, representing a typical abnormal river state, which disappeared after 2021.

[0026] Based on the evaluation results of this embodiment, it was calculated that the section of the Yellow River with the most severe local swinging in the lower reaches was a certain control and guidance project section. By comparing remote sensing images and river morphology monitoring data, the abnormal river morphology at this location was clearly developed. After a long period of heavy flooding during the 2020 flood season, the abnormal river morphology disappeared, verifying the engineering guidance value of this method.

[0027] Example 2 The river section to be evaluated in this embodiment is the same as in Embodiment 1. The method described in this invention is used to analyze the deviation between the pre-flood measured river morphology of the Yellow River and the guiding line of the downstream channel regulation project, in order to verify the conformity between the main river morphology line and the planned guiding line. The specific implementation steps are as follows: S1, Preprocessing of basic river channel data: (1) Data sources: Baseline data: Vector data of the planned traverse line of the river section to be evaluated; Comparison data: Vector data of the main channel line measured before the flood season in 2020 in Example 1; Figure 3 The image shows a comparison between the measured main channel line of the Yellow River and the planned guiding line before the flood season in the lower reaches of the Yellow River in 2020.

[0028] (2) Outlier removal: The sliding window 3σ criterion is used to remove shallow water and noise outliers in the measured mainstream line; (3) Coordinate system one: The two sets of vector data are uniformly converted into the general Gauss-Kruger 3° zone plane projection coordinate system of the Yellow River Basin, with a central longitude of 114°; (4) Unified sequence direction: The planned traverse line and the measured main traverse line are both adjusted to be arranged in a unidirectional orderly manner from upstream to downstream to ensure that the start and end points of the sequence are completely corresponding; (5) Adaptive equal-interval resampling: For this wandering river section, resampling is carried out at equal intervals of 100m to obtain the planned traverse sequence P and the measured main traverse sequence Q. The two sequences contain 3542 and 3404 discrete points, respectively.

[0029] S2, Friesian distance calculation: Solving the original Friesian distance d using a dynamic programming algorithm with ordered downstream matching. F =3414m, normalized calculation yields: d F−norm =0.011 indicates that the measured overall river morphology is basically similar to the planned traverse line deviation.

[0030] S3, Two-way Hausdorff distance calculation: First, traverse all points of the planned traverse line to match the measured main line; then, traverse the measured main line in reverse to match the planned traverse line. After comparing extreme values ​​in both directions, the final Hausdorff distance d of the river segment is obtained. H =3415m, which represents the maximum lateral deviation between the measured mainstream line and the planned guiding line in 2020, corresponding to the risk of severe local swaying.

[0031] S4, Two-Dimensional Collaborative Joint Evaluation: Combining the two-dimensional matrix joint evaluation system and matrix grading standard, the normalized Frescher distance of 0.011 > 0.01 indicates that the river morphology and the planned guiding line in 2020 were generally similar. The two-way Hausdorf distance of 3415m is greater than 3000m, corresponding to severe local river channel swings. Comparison with remote sensing images reveals that this swing location is near a certain control project, which is a typical abnormal river morphology.

[0032] Based on the evaluation results of this embodiment, dangerous sections that deviate from the planned guiding line were accurately located, and the flood control risks that may be caused by river deviation were effectively identified.

[0033] Furthermore, it can be seen that the Fraser distance and the bidirectional Hausdorff distance calculated in this study are equal. This is because, mathematically, the Hausdorff distance is a natural lower bound of the Fraser distance. The Fraser distance represents a subset of all possible pairings in the Hausdorff distance (which must follow the rule of monotonicity and non-returning), therefore the Fraser distance can only be greater than or equal to the Hausdorff distance. Since the riverbanks compared from different years all follow the same direction from upstream to downstream, with the starting and ending points strictly corresponding to the same control section, and monotonically increasing along the flow direction without self-intersections or hairpin bends, they satisfy the curve characteristic that the Fraser distance and Hausdorff distance are equal. Therefore, the calculation results are reasonable.

Claims

1. A method for evaluating the stability of wandering river channels based on trajectory similarity algorithm, characterized in that, Includes the following steps: S1, River vector data preprocessing: Obtain river vector data of different periods of the wandering river section to be evaluated, remove outliers, unify coordinate system and sequence direction, and resample at equal intervals to obtain the discrete point sequence coordinates [X,Y] arranged unidirectionally from upstream to downstream of the river at different periods, ensuring that the start and end points of the sequence correspond to the starting and ending sections of the wandering river section to be evaluated, respectively. S2, Discrete Fraser distance calculation: Based on any two pre-processed sets of unidirectionally ordered discrete point sequences, the discrete Fraser distance is calculated using a dynamic programming algorithm, and then normalized by combining it with the actual length of the benchmark river channel to obtain the normalized Fraser distance, which is used to characterize the evolution process and laws of the overall topological morphology of the river. S3, Bidirectional Hausdorff distance calculation: Based on any two preprocessed sets of unidirectionally ordered discrete point sequences, calculate the bidirectional directed Hausdorff distance, and take the maximum value of the two directed distances as the final Hausdorff distance. S4, Two-Dimensional Collaborative Joint Evaluation: Based on Normalized Fréchet Distance Distance between two-way Hausdorf A two-dimensional matrix joint evaluation system is constructed, which, combined with the pre-set river stability evaluation and grading standards, outputs the final river status evaluation results.

2. The method for evaluating the stability of wandering river channels based on trajectory similarity algorithm according to claim 1, characterized in that: In step S1, the river vector line data is either the main river channel vector data or the planning and guiding line vector data of the river regulation project. The data sources include river morphology survey reports, remote sensing impact interpretation, large cross-section observation data, and approved river regulation line data. The coordinate system is the Gauss-Kruger 3° plane projection coordinate system, which eliminates the influence of geographic coordinate projection on distance calculation.

3. The method for evaluating the stability of wandering river channels based on trajectory similarity algorithm according to claim 1, characterized in that: In step S1, outlier removal is performed using a sliding window. The criteria are as follows: using 5 consecutive sampling points as a sliding window, calculate the mean and standard deviation of the horizontal offset of the points within the window, and remove outliers that deviate from the mean by more than a preset threshold.

4. The method for evaluating the river morphology stability of wandering rivers based on trajectory similarity algorithm according to claim 1, characterized in that: In step S2 The formula for calculating the normalized Fréchet distance is as follows: In the formula, The actual length of the river channel vector line during the baseline period is normalized to a dimensionless value of 0-1. It is the discrete Friesian distance between the preprocessed, unidirectionally ordered discrete point sequence P used as the benchmark and the preprocessed, unidirectionally ordered discrete point sequence Q used for comparison.

5. The method for evaluating the stability of wandering river channels based on trajectory similarity algorithm according to claim 1, characterized in that: In step S3, the formula for calculating the bidirectional directed Hausdorff distance is: Where P is the preprocessed baseline period unidirectional ordered discrete point sequence, and Q is the preprocessed comparison period unidirectional ordered discrete point sequence. The directed Hausdorff distance from P to Q is calculated using the following formula: In the formula, The Euclidean distance between the two points; Let Q be the directed Hausdorff distance from P; The calculation formula is: In the formula, This is the Euclidean distance between two points.

6. The method for evaluating the stability of wandering river channels based on trajectory similarity algorithm according to claim 1, characterized in that: In step S4, the river state classification standard of the two-dimensional matrix joint evaluation system is specifically as follows: 。 7. The method for evaluating the stability of wandering river channels based on trajectory similarity algorithm according to claim 1, characterized in that: The two sets of unidirectionally ordered discrete point sequences are measured mainstream river data of the same river segment at different times, used for analyzing the interannual evolution of river morphology; or the two sets of unidirectionally ordered discrete point sequences are mainstream river vector data and planned traverse line data of the same river segment, used for analyzing the degree of deviation between river morphology and planned traverse line.

Images