CN116012741A - Water and Soil Erosion Monitoring System for High-Voltage Transmission Lines - Google Patents

Abstract

The invention discloses a high-voltage transmission line water and soil loss monitoring system, which belongs to the technical field of water and soil loss monitoring. The system includes: a data acquisition module: used to collect water and soil loss monitoring images along the high-voltage transmission line in the target area through a drone, Form a monitoring image sequence; preprocessing module: used to preprocess the monitoring images in the monitoring image sequence; image registration module: use the SIFT algorithm to perform feature extraction on the preprocessed monitoring image sequence, using the chaos game optimization algorithm The optimal feature point matching is performed on two adjacent monitoring images; the image fusion module is used for image transformation, and the weighted average method is used to perform image fusion on two adjacent monitoring images to obtain a panoramic monitoring image of water and soil erosion. The present invention collects monitoring images of water and soil loss by unmanned aerial vehicles, and uses an improved chaos game optimization algorithm to match the optimal feature points of the images, so that the panoramic monitoring of water and soil loss on high-voltage transmission lines can be quickly realized.

Description

Soil and Water Loss Monitoring System for High Voltage Transmission Lines

Technical Field

The invention belongs to the technical field of soil and water loss monitoring, and in particular relates to a soil and water loss monitoring system for a high-voltage power transmission line.

Background Art

With the construction of power infrastructure, the implementation of some power transmission and transformation projects on high-voltage transmission lines has brought about a large amount of vegetation and land damage, which has damaged the water and soil resources along the line to a certain extent and caused serious soil erosion. In order to reduce the man-made soil erosion caused by power construction, it is necessary to formulate a soil erosion monitoring plan and corresponding soil and water conservation measures to provide a scientific basis for the green development of power construction.

High-voltage transmission lines are mainly composed of iron towers and transmission lines. The characteristics of soil erosion caused by the erection of high-voltage transmission lines are complex. As a linear project, it has the characteristics of large spatial span, scattered disturbance points, and diversified landform types and soil erosion types involved. In addition to the soil erosion during the construction of the tower foundation, felling, excavation, backfilling and other projects through which the line passes, after the completion of the project, the high-voltage transmission line will also suffer from secondary soil erosion caused by meteorological factors, geological factors, etc. during long-term use. For example, the invention patent with publication number CN114219678A discloses a method for monitoring and early warning of soil erosion in the construction of overhead line tower foundations. It analyzes the working condition indicators related to soil erosion for the processes and contents related to tower foundation construction in the construction of power transmission and transformation line projects, and calculates the efficacy coefficient values of different construction projects for soil erosion monitoring and early warning. It realizes the monitoring and early warning of soil erosion caused by tower foundation construction, but it cannot fully monitor and analyze the soil erosion of the entire transmission line. The invention patent with publication number CN115239566A discloses a soil erosion monitoring method based on ultra-high voltage transmission lines, which collects planar soil erosion data by sequentially setting soil erosion sensors along the transmission line, generates corresponding images, and performs image matching and splicing to obtain linear soil erosion data of the transmission line. It can improve the degree of monitoring automation and obtain soil erosion data of the complete transmission line, but due to the large spatial span of high-voltage transmission lines and the complex environment along the line, the construction efficiency of manually burying a large number of sensors is low, and sensor damage and monitoring failure are prone to occur, and subsequent maintenance is difficult, which is not conducive to rapid and stable soil erosion monitoring.

Summary of the invention

In view of this, the present invention proposes a soil erosion monitoring system for a high-voltage transmission line, which is used to solve the problem that the existing soil erosion monitoring scheme for a high-voltage transmission line cannot realize rapid monitoring of the entire line.

The present invention provides a soil and water loss monitoring system for a high-voltage transmission line, the system comprising:

Data acquisition module: used to collect soil and water loss monitoring images along the high-voltage transmission lines in the target area through drones to form a monitoring image sequence;

Preprocessing module: used for preprocessing monitoring images in the monitoring image sequence;

Image registration module: It is used to extract features from the preprocessed monitoring image sequence through the SIFT algorithm, and to match the optimal feature points of two adjacent monitoring images using the chaotic game optimization algorithm with the goal of minimizing the Euclidean distance between the two adjacent monitoring images;

Image fusion module: used to perform image transformation based on the optimal feature point matching results of two adjacent monitoring images, and use the weighted average method to fuse the two adjacent monitoring images in turn to obtain a panoramic monitoring image of soil and water loss on the high-voltage transmission line in the target area.

On the basis of the above technical solution, preferably, in the data acquisition module,

The quality of the soil and water loss monitoring images is inspected according to the acquisition sequence, and images that meet the image stitching quality requirements are screened to form a monitoring image sequence, in which two adjacent monitoring images overlap.

On the basis of the above technical solution, preferably, in the preprocessing module, the preprocessing specifically includes:

The images in the monitoring image sequence are denoised and distortion corrected respectively.

On the basis of the above technical solution, preferably, the use of the chaotic game optimization algorithm to match the optimal feature points of two adjacent monitoring images specifically includes:

Determine the overlapping area of two adjacent monitoring images, and determine the search space range of feature points and the dimension of the solution according to the range of the overlapping area;

Randomly initialize candidate solutions of the chaos game optimization algorithm within the search space;

The fitness function is designed with the goal of minimizing the Euclidean distance between two adjacent monitoring images;

Calculate the fitness value of each current candidate solution, and record the global optimal solution X _bt , suboptimal solution X _bs and worst solution X _ws ;

For each candidate solution _Xi , multiple candidate solutions are randomly selected within the search space, and the mean update rule of the vector weighted average algorithm is introduced to calculate the weighted mean _MRi of multiple candidate solutions;

For each candidate solution _Xi , a temporary triangle is determined using the position of the current candidate solution _Xi , the global optimal solution _Xbt and the weighted mean _MRi of multiple candidate points;

For each temporary triangle, four seed points are generated for position update;

Calculate the fitness value of the seed point and update the global optimal solution;

Determine whether the maximum number of iterations is met. If so, output the global optimal solution. Otherwise, re-iterate the calculation.

On the basis of the above technical solution, preferably, for each temporary triangle, generating four seed points for position update specifically includes:

On the basis of the mean update rule of the weighted average algorithm, an acceleration rule based on direction judgment is introduced to generate the first seed point position according to the following formula:

；

;

，f为适应度函数，sign为符号函数，用于判断较优解的方向；

是随机生成的矩阵，用于模拟种子的运动位置限制，

、

表示1或2的随机整数；in,

, f is the fitness function, sign is the sign function, which is used to determine the direction of the better solution;

is a randomly generated matrix used to simulate the movement position constraints of the seed.

,

A random integer representing 1 or 2;

The second seed point position is generated according to the following formula:

；

;

The third seed point position is generated according to the following formula:

；

;

The fourth seed point position is generated according to the following formula:

；

;

为点对点乘法，

为莱维随机搜索函数，服从参数为

的莱维分布。Among them, R is the step size control value of D dimension,

is point-to-point multiplication,

is the Levy random search function, which is subject to the parameter

The Levy distribution.

On the basis of the above technical solution, preferably, the mean update rule of the vector weighted average algorithm is introduced to calculate the weighted mean MR _i of multiple candidate solutions as follows:

；

;

为三个随机候选解，r为[0,0.5]内的随机数，

，t为当前迭代次数，T为最大迭代次数，

为0~2之间的随机数，

为WM1的三个加权函数，

为WM2的三个加权函数。in,

are three random candidate solutions, r is a random number in [0,0.5],

, t is the current iteration number, T is the maximum iteration number,

is a random number between 0 and 2.

are the three weighting functions of WM1,

are the three weighting functions of WM2.

On the basis of the above technical solution, preferably, the expression of the fitness function is:

；

;

表示d_k的第m维分量，m=1,2,...,M为特征向量的维度，k=1,2,…,n，n为参考图像中待匹配的特征点总数，d_j为浮动图像中的第j个特征点的特征向量，λ_k为权重系数。Where d _k is the feature vector of the kth feature point in the reference image,

represents the m-th dimension component of _dk , where m=1,2,…,M is the dimension of the feature vector, k=1,2,…,n, n is the total number of feature points to be matched in the reference image, _dj is the feature vector of the j-th feature point in the floating image, and _λk is the weight coefficient.

On the basis of the above technical solution, preferably, in the image fusion module, performing image transformation based on the optimal feature point matching results of two adjacent monitoring images specifically includes:

The RANSAC algorithm is used to eliminate the false matching points and obtain the accurate matching points. The image transformation matrix is calculated based on the accurate matching points. Any one of the two adjacent monitoring images is used as the reference image, and the other image is transformed onto the reference image according to the image transformation matrix.

Compared with the prior art, the present invention has the following beneficial effects:

1) The present invention collects soil and water loss monitoring images along the high-voltage transmission line in the target area through a drone to form a monitoring image sequence. By using a chaotic game optimization algorithm to match the optimal feature points of two adjacent monitoring images, a panoramic soil and water loss monitoring image can be quickly spliced to achieve full-line soil and water loss monitoring on the high-voltage transmission line;

2) The present invention introduces the mean update rule of the vector weighted average algorithm into the chaotic game optimization algorithm to calculate the weighted mean MR _i of multiple candidate solutions, thereby improving the rationality of the position of the temporary triangle, reducing the shortcoming of the excessive jump in the position of the temporary triangle caused by the original chaotic game optimization algorithm directly determining the average value according to the average values of multiple randomly selected candidate solutions, and improving the algorithm search capability;

3) In the process of generating seed points for each temporary triangle, the present invention introduces an acceleration rule based on direction judgment to update the position of the first seed point on the basis of the mean update rule of the weighted average algorithm, thereby accelerating the convergence speed of the algorithm and improving the registration speed. A position update method based on Levy flight is used to replace the random movement of the fourth seed point to avoid falling into a local optimal solution.

BRIEF DESCRIPTION OF THE DRAWINGS

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

FIG1 is a structural diagram of a soil and water loss monitoring system for high-voltage power transmission lines according to the present invention.

DETAILED DESCRIPTION

The following will be combined with the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described implementation of the high-voltage transmission line soil and water loss monitoring system is only a part of the implementation of the present invention, not all of the implementations. Based on the implementation of the present invention, all other implementations obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

Please refer to FIG. 1 . The present invention provides a soil erosion monitoring system for a high-voltage transmission line. The system includes a data acquisition module, a preprocessing module, an image registration module and an image fusion module.

The data acquisition module is used to collect soil and water loss monitoring images along the high-voltage transmission lines in the target area through drones to form a monitoring image sequence.

Specifically, a drone equipped with a camera can be used to fly along the high-voltage transmission lines in the target area to collect monitoring videos along the high-voltage transmission lines, or images along the high-voltage transmission lines can be collected according to a preset period to obtain soil and water loss monitoring images along the high-voltage transmission lines.

Since the images taken by drones during flight are affected by weather, shooting angles, etc., the images taken are uneven and need to be screened. The quality of soil and water loss monitoring images can be tested in the order of acquisition, and images that meet the image stitching quality requirements can be screened to form a monitoring image sequence. The overlap between two adjacent monitoring images in the monitoring image sequence is required to be within a preset range. Image quality detection indicators such as clarity, light quantization value, and effective area ratio can also be set.

The preprocessing module is used to preprocess the monitoring images in the monitoring image sequence.

Due to the influence of internal and external factors and the wireless transmission process, the quality of the images collected by the drone will degrade to varying degrees. Therefore, it is necessary to perform denoising and distortion correction on the images in the monitoring image sequence respectively.

The image registration module is used to extract features from the preprocessed monitoring image sequence through the SIFT algorithm, and to minimize the Euclidean distance between two adjacent monitoring images, and to use the chaotic game optimization algorithm to perform optimal feature point matching on the two adjacent monitoring images.

During the aerial photography process, drones are affected by the flight altitude, airflow and camera viewing angle, resulting in different scales of the captured target objects. The SIFT algorithm can handle the feature matching problems between two images under the conditions of translation, rotation, scale change and illumination change, and has a relatively stable matching ability for perspective change and affine change to a certain extent. Therefore, the present invention describes the image features through the SIFT algorithm, and the image registration based on the SIFT algorithm is based on the entire image, with many feature points and a large dimension of each feature point, so there are problems of large feature matching data volume and long time consumption. Therefore, the present invention adopts a chaotic game optimization algorithm to match the optimal feature points of two adjacent monitoring images to improve the speed and accuracy of image matching.

The image registration module specifically includes a feature extraction unit and a registration optimization unit.

The feature extraction unit is used to extract features from the preprocessed monitoring image sequence through the SIFT algorithm.

Specifically, the use of SIFT algorithm to extract image features generally includes four steps: ① constructing a scale space and detecting extreme points; ② determining candidate feature points based on the detected extreme points; ③ determining the direction of the feature points; ④ generating feature vector descriptors. Assuming that in the SIFT algorithm, the neighborhood size is 4×4, and the gradient values are distributed in 8 directions, such a feature point shares 128 data descriptions, and finally generates a 128-dimensional feature vector descriptor, then the dimension of the feature point M=128.

The registration optimization unit is used to match the optimal feature points of two adjacent monitoring images by using a chaotic game optimization algorithm with the goal of minimizing the Euclidean distance between the two adjacent monitoring images.

Any one of the two adjacent monitoring images is taken as the reference image, and the other image is taken as the floating image. Image registration is to match the feature points between the reference image and the floating image. The present invention takes the reference image as a benchmark and adopts a chaotic game optimization algorithm to select a group of feature points from the floating image so that it can be quickly matched with each feature point to be matched in the reference image.

The execution process of the registration optimization unit specifically includes the following steps:

S1. Determine the overlapping area of two adjacent monitoring images, determine the search space range [X _min , X _max ] of feature points and the dimension D of the solution according to the range of the overlapping area, and initialize various parameters of the chaotic game optimization algorithm.

The reference image and the floating image are matched based on the overlapping area, so the search space range of the feature points can be determined based on the pixel range of the overlapping area. The dimension D of the solution is the same as the number n of feature points to be matched in the reference image.

S2. Randomly initialize the candidate solutions X=[X ₁ ,X ₂ ,…, _Xi ,…,X _N ] of the chaotic game optimization algorithm within the search space, where i=1,2,…,N, N is the number of candidate solutions, and one candidate solution corresponds to the position of a set of feature points in the floating image.

X _i =X _i,min +rand*(X _i,max -X _i,min )

_Xi is the i-th candidate solution, Xi _,max and Xi _,min are the upper and lower limits of _Xi ’s search space, respectively.

S3. Calculate the fitness value of each current candidate solution, and record the global optimal solution X _bt , the suboptimal solution X _bs , and the worst solution X _ws .

The fitness function is designed with the goal of minimizing the Euclidean distance between two adjacent monitoring images. The expression of the fitness function is:

；

;

表示d_k的第m维分量，m=1,2,...,M为特征向量的维度，k=1,2,…,n，n为参考图像中待匹配的特征点总数， d_j为浮动图像中的第j个特征点的特征向量，即当前候选解的某一个位置分量对应的特征点的特征向量，λ_k为权重系数。Where d _k is the feature vector of the kth feature point in the reference image,

represents the m-th dimension component of _dk , where m=1,2,…,M is the dimension of the feature vector, k=1,2,…,n, where n is the total number of feature points to be matched in the reference image, _dj is the feature vector of the j-th feature point in the floating image, that is, the feature vector of the feature point corresponding to a certain position component of the current candidate solution, and _λk is the weight coefficient.

S4. For each candidate solution _Xi , randomly select multiple candidate solutions from the search space, and introduce the mean update rule of the vector weighted average algorithm to calculate the weighted mean _MRi of the multiple candidate solutions.

，引入向量加权平均算法的均值更新规则MeanRule，结合全局最优解X_bt、次优解X_bs和最差解X_ws计算这三个随机候选解的加权均值MR_i，MR_i的计算公式具体为：Specifically, for each candidate solution _Xi , three candidate solutions are randomly selected from the search space.

, the mean update rule MeanRule of the vector weighted average algorithm is introduced, and the weighted mean MR _i of the three random candidate solutions is calculated by combining the global optimal solution X _bt , the suboptimal solution X _bs and the worst solution X _ws . The calculation formula of MR _i is as follows:

；

;

，t为当前迭代次数，T为预先设定的最大迭代次数，

为0~2之间的随机数。Where r is a random number in [0,0.5],

, t is the current iteration number, T is the preset maximum iteration number,

A random number between 0 and 2.

为WM1的三个加权函数，表达式为：

are the three weighted functions of WM1, and the expressions are:

；

;

为WM2的三个加权函数，表达式为：

are the three weighted functions of WM2, and the expressions are:

；

;

S5. For each candidate solution _Xi , a temporary triangle is determined using the positions of the current candidate solution _Xi , the global optimal solution _Xbt and the weighted mean _MRi of multiple candidate points.

S6. For each temporary triangle, four seed points are generated to update the position.

Step S6 specifically includes the following sub-steps:

S61. Based on the mean update rule of the weighted average algorithm, an acceleration rule based on direction judgment is introduced to generate the first seed point position:

;

;

，f为适应度函数，sign为符号函数，用于判断较优解的方向；当较优解在中心点d₀的左侧，则使第一种子点向左移动，否则向右移动，δ_t是步长调节因子，

是D维的单位方向向量，α_i是随机生成的矩阵，用于模拟种子的运动位置限制，β_i、γ_i表示值为1或2的随机整数；in,

, f is the fitness function, sign is the sign function, which is used to determine the direction of the better solution; when the better solution is on the left side of the center point d ₀ , the first seed point is moved to the left, otherwise it is moved to the right, δ _t is the step size adjustment factor,

is a D-dimensional unit direction vector, α _i is a randomly generated matrix used to simulate the movement position restriction of the seed, β _i and γ _i represent random integers with values of 1 or 2;

S62, generating the second seed point position:

;

;

S63, generating the third seed point position:

;

;

S64, generating the fourth seed point position:

;

;

为点对点乘法，

为莱维随机搜索函数，服从参数为

的莱维分布，其表达式为：Among them, R is the step size control value of D dimension,

is point-to-point multiplication,

is the Levy random search function, which is subject to the parameter

The Levy distribution is expressed as:

;

;

Among them, is the standard gamma function, and u and v both obey the normal distribution.

S7. Calculate the fitness value of the seed point and update the global optimal solution.

Each candidate solution _Xi can determine a temporary triangle. Each temporary triangle corresponds to four seed points. The fitness values of the seed points are calculated respectively, and compared and sorted with the fitness values of each candidate solution in step S3 to update the global optimal solution. The position of the corresponding candidate solution is updated with the optimal position of the four seed points corresponding to each candidate solution _Xi to obtain a new candidate solution position.

S8. Determine whether the maximum number of iterations is met. If so, output the global optimal solution. Otherwise, return to step S3 and iterate again.

The present invention adopts a chaotic game optimization algorithm to perform registration optimization in the image registration module, and introduces the mean update rule of the vector weighted average algorithm to calculate the weighted mean MR _i of multiple candidate solutions, thereby improving the rationality of the position of the temporary triangle, reducing the deficiency of excessive position jump of the temporary triangle caused by the original chaotic game optimization algorithm directly determining the average value according to the average value of multiple randomly selected candidate solutions, and improving the algorithm search capability. At the same time, in the process of generating seed points for each temporary triangle, on the basis of the mean update rule of the weighted average algorithm, an acceleration rule based on direction judgment is introduced to update the position of the first seed point, thereby accelerating the convergence speed of the algorithm, and adopting a position update method based on Levy flight to replace the random movement of the fourth seed point, thereby avoiding falling into a local optimal solution.

Image fusion module: used to perform image transformation based on the optimal feature point matching results of two adjacent monitoring images, and use the weighted average method to fuse the two adjacent monitoring images in turn to obtain a panoramic monitoring image of soil and water loss on the high-voltage transmission line in the target area.

The existence of mismatching points will affect the quality of image stitching. Therefore, before image fusion, the RANSAC algorithm is used to eliminate the mismatching points and obtain the precise matching points. The image transformation matrix is calculated based on the precise matching points. Any one of the two adjacent monitoring images is used as the reference image, and the floating image is transformed to the reference image according to the image transformation matrix.

The present invention collects soil and water loss monitoring images along the high-voltage transmission line in the target area through an unmanned aerial vehicle to form a monitoring image sequence. By using a chaotic game optimization algorithm to match the optimal feature points of two adjacent monitoring images, a panoramic monitoring image of soil and water loss can be quickly spliced to achieve full-line monitoring of soil and water loss on the high-voltage transmission line. Compared with the monitoring method of deploying sensors, the present invention can improve the degree of monitoring automation and obtain soil and water loss data of the complete transmission line, and is suitable for high-voltage transmission lines with large spatial spans and complex environments along the line. It has high mobility and is conducive to rapid and stable soil and water loss monitoring. In addition, the present invention abandons the conventional image registration method in the process of unmanned aerial vehicle image splicing, and improves the chaotic game search optimization algorithm by introducing the mean update rule of the weighted average algorithm and the acceleration rule based on direction judgment of the longhorn beard optimization algorithm and the Levy flight optimization strategy to improve the speed and accuracy of image registration, and finally obtain a panoramic image with fast splicing speed and high imaging quality, which is convenient for soil and water loss monitoring and analysis of the entire high-voltage transmission line.

The present invention also discloses an electronic device, comprising: at least one processor, at least one memory, a communication interface and a bus; wherein the processor, memory, and communication interface communicate with each other through the bus; the memory stores program instructions that can be executed by the processor, and the processor calls the program instructions to implement the aforementioned system of the present invention.

The present invention also discloses a computer-readable storage medium, wherein the computer-readable storage medium stores computer instructions, wherein the computer instructions enable the computer to implement all or part of the functions of the system described in the embodiment of the present invention. The storage medium includes: a USB flash drive, a mobile hard disk, a read-only memory ROM, a random access memory RAM, a magnetic disk or an optical disk, and other media that can store program codes.

The system embodiment described above is merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be distributed to multiple network units. A person skilled in the art may select some or all of the modules according to actual needs to achieve the purpose of the solution of this embodiment without creative work.

The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims (8)

1. A high voltage transmission line soil erosion monitoring system, the system comprising:

and a data acquisition module: the method comprises the steps that water and soil loss monitoring images along a high-voltage transmission line in a target area are collected through an unmanned aerial vehicle, and a monitoring image sequence is formed;

and a pretreatment module: the method comprises the steps of preprocessing monitoring images in a monitoring image sequence;

an image registration module: the method comprises the steps of performing feature extraction on a preprocessed monitoring image sequence through a SIFT algorithm, and performing optimal feature point matching on two adjacent monitoring images by adopting a chaotic game optimization algorithm with the minimum Euclidean distance between the two adjacent monitoring images as a target;

and an image fusion module: the method is used for carrying out image transformation based on the optimal feature point matching result of two adjacent monitoring images, and carrying out image fusion on the two adjacent monitoring images in sequence by adopting a weighted average method to obtain the panoramic monitoring image of water and soil loss on the high-voltage transmission line in the target area.

2. The system for monitoring water and soil loss of a high-voltage transmission line according to claim 1, wherein in the data acquisition module, quality detection is performed on water and soil loss monitoring images according to an acquisition sequence, an image forming monitoring image sequence meeting the image splicing quality requirement is screened, and two adjacent monitoring images in the monitoring image sequence are overlapped.

3. The water and soil loss monitoring system of a high-voltage transmission line according to claim 1, wherein the preprocessing module specifically comprises:

and respectively denoising and correcting distortion of images in the monitoring image sequence.

4. The system for monitoring water and soil loss of a high-voltage transmission line according to claim 1, wherein the method for performing optimal feature point matching on two adjacent monitoring images by adopting a chaotic game optimization algorithm specifically comprises:

determining overlapping areas of two adjacent monitoring images, and determining search space ranges and dimensionality of the solution of the feature points according to the ranges of the overlapping areas;

randomly initializing candidate solutions of a chaotic game optimization algorithm in a search space range;

the Euclidean distance between two adjacent monitoring images is used as a target design fitness function;

calculating the fitness value of each candidate solution at present, and recording the global optimal solution X _bt Suboptimal solution X _bs And worst solution X _ws ；

For each candidate solution X _i Randomly selecting a plurality of candidate solutions in the search space range, and introducing a mean value updating rule of a vector weighted average algorithm to calculate a weighted mean MR of the plurality of candidate solutions _i I=1, 2, …, N is the total number of candidate solutions;

for each candidate solution X _i With current candidate solution X _i Global optimal solution X _bt And a weighted average MR of the plurality of candidate points _i A temporary triangle is determined by the position of the (a);

for each temporary triangle, respectively generating four seed points for position updating;

calculating the fitness value of the seed points and updating the global optimal solution;

judging whether the maximum iteration times are met, if so, outputting a global optimal solution, otherwise, carrying out iterative computation again.

5. The system for monitoring water and soil loss of a high voltage transmission line according to claim 4, wherein for each temporary triangle, generating four seed points for position update respectively comprises:

introducing an acceleration rule based on direction judgment on the basis of a mean value updating rule of a weighted average algorithm, and generating a first seed point position according to the following formula:

；/>

wherein ,

，fin order to adapt the function of the degree of adaptation,signis a sign function and is used for judging the direction of a better solution; delta _t Is a step size adjustment factor,/->

Is the unit direction vector of D dimension, alpha _i Is a randomly generated matrix for simulating the motion position limitation of seeds, beta _i 、γ _i Representing a random integer having a value of 1 or 2;

generating a second seed point location according to the formula:

；

generating a third seed point location according to the formula:

；

generating a fourth seed point location according to the formula:

；

wherein R is the step control quantity of D dimension,

for point-to-point multiplication, +.>

For the Lewy random search function, the obeying parameter is +.>

Is a distribution of the Lewy of (C).

6. The system for monitoring water and soil erosion of high voltage transmission line according to claim 4, wherein the mean update rule incorporating the vector weighted average algorithm calculates weighted mean MR of a plurality of candidate solutions _i The method comprises the following steps:

；

wherein ,

for three random candidate solutions, r is [0,0.5]The random number in the random number is used for the random number,

t is the current iteration number, T is the maximum iteration number, +.>

Is a random number between 0 and 2, < >>

Is three weighting functions of WM1,

three weighting functions for WM 2.

7. The system for monitoring water and soil loss of a high voltage transmission line according to claim 4, wherein the fitness function has an expression as follows:

；

wherein ,d_k For the feature vector of the kth feature point in the reference image, m=1, 2..m is the dimension of the feature vector, k=1, 2, …, n, n is the total number of feature points to be matched in the reference image, d _j Lambda is the feature vector of the j-th feature point in the floating image _k Is a weight coefficient; the reference image is any one image of two adjacent monitoring images, and the other image is a floating image.

8. The system for monitoring water and soil loss of a high-voltage transmission line according to claim 1, wherein in the medium-image fusion module, performing image transformation based on an optimal feature point matching result of two adjacent monitoring images specifically comprises:

and eliminating the mismatching points through a RANSAC algorithm to obtain accurate matching points, calculating an image transformation matrix according to the accurate matching points, taking any one image of two adjacent monitoring images as a reference image, and transforming the other image onto the reference image according to the image transformation matrix.

Images