CN-122015708-A - Power distribution tower pole deformation monitoring system based on visual detection

CN122015708ACN 122015708 ACN122015708 ACN 122015708ACN-122015708-A

Abstract

The invention discloses a power distribution tower pole deformation monitoring system based on visual detection, which relates to the technical field of power transmission and transformation monitoring, and specifically comprises monitoring of debris flow susceptibility, monitoring of weather disasters, monitoring of landslide disasters, access of monitoring equipment, establishment of an edge algorithm model and processing of Beidou data. According to the invention, access and management of monitoring equipment and intelligent terminals are completed by deploying the towers and the surrounding environment monitoring, analyzing and early warning system thereof, the intelligent terminals, cameras and other sensor equipment are deployed on the towers, acquisition of health data and image data of the towers is realized, the intelligent terminals are added with Beidou modules, beidou communication capacity is realized, the monitoring algorithm model and the data and image compression algorithm model are deployed on the intelligent terminals, analysis and compression of the data and the images are realized, and the monitoring and early warning of remote towers and surrounding environments thereof are realized by uploading the Beidou modules to a monitoring center.

Inventors

Ren Yizhang
WU SHIFAN
WANG BEN
YU ZONGBO
GUO QIANG
LI CHANGBO
ZHU XIAOMING
Zhen Gen
JIAO YAQIN
SHA PENG
ZHAO LONGQIAN
SUN JIANQUAN
SHE QIANG
BAO YULIN
AN YACHENG
YAO WEI
HUANG LIFEI
WANG HAOYU

Assignees

国网青海省电力公司西宁供电公司
国网山西省电力有限公司太原供电分公司
国网湖南省电力有限公司长沙供电分公司

Dates

Publication Date: 20260512
Application Date: 20260325

Claims (10)

1. The power distribution tower pole deformation monitoring system based on visual detection is characterized by comprising monitoring of debris flow susceptibility, monitoring of weather disasters, monitoring of landslide disasters, access of monitoring equipment, establishment of an edge algorithm model and processing of Beidou data; The monitoring of the debris flow susceptibility is specifically to evaluate the debris flow susceptibility by adopting a machine learning algorithm, and the monitoring of the debris flow susceptibility is divided into four processes, namely, pretreatment of data, construction of a susceptibility evaluation model, partitioning of the debris flow susceptibility according to model results and evaluation of model performance; The monitoring of the meteorological disaster specifically refers to the study of annual change according to precipitation data; the monitoring of landslide disasters specifically refers to training a model for identifying landslide by using a network of YOLO V5 in deep learning; The access of the monitoring equipment is specifically based on a cloud edge end integrated architecture, and millions of monitoring equipment access capacity, monitoring data visual analysis and early warning capacity and remote operation and maintenance management capacity are built; The establishment of the edge algorithm model specifically refers to that a manager remotely updates the algorithm model according to monitoring requirements of different regions and different periods, and research and development of the monitoring center algorithm model and training results are applied to the edge side; The Beidou data processing specifically refers to the fact that the Beidou data processing method is used for comprehensively improving the efficiency of data transmission of monitoring data by satellite communication through a data compression algorithm.
2. The visual detection-based power distribution tower pole deformation monitoring system according to claim 1, wherein the monitoring of the susceptibility to debris flow comprises the selection and classification of stone flow influence factors, the analysis of the importance of the influence factors and the establishment of influence factor models.
3. The visual detection-based power distribution tower pole deformation monitoring system is characterized in that the selection classification of the debris flow influencing factors adopts a natural break method to determine an optimal classification interval, and the method specifically comprises the steps of calculating the variances of a group of data X= { X 1 , x 2 ,…,x n }（x 1 <=x 2 <=…<=x n ) as SDAM, assuming that the group of data is divided into M groups, calculating the variances of each group by knowing any group interval Yi={x 1 , y 1 , y 2 ,…, y M-1 , x n }（x 1 <y 1 <y 2 <…<y M-1 <x n ）,, summing the variances by (x 1 ,y 1 ), (y 1 ,y 2 ), (y 2 ,y 3 ),…,(y M-2 ,y M-1 ), (y M- 1,x n ), to obtain the sum of the variances of all the groups, calculating the sum as SDCM, and calculating the corresponding SDCMi for different classification intervals Yi; Selecting the minimum SDCMi corresponding classification interval, namely the optimal classification interval; The calculation formulas of SDAM and SDCM are respectively as follows: Where Nj represents the number of samples of class j, The minimum SDCM value corresponds to the maximum variance fitting goodness-of-measure GVF value, and the GVF calculation formula is as follows: wherein the GVF value range is [0,1].
4. The visual detection-based power distribution tower pole deformation monitoring system according to claim 2, wherein the analysis of the importance of the influence factors is specifically that the information of geological background and debris flow development characteristics is analyzed to qualitatively select debris flow influence factors; evaluating the importance of the selected influence factors by adopting a quantitative method; selecting the most appropriate influence factors by a means of combining a qualitative method with a quantitative method; The specific operation of the analysis of the importance of the influence factors is that) the information entropy, the conditional entropy and the information gain value are obtained through calculation and are used for calculating the information gain rate, For a set of data y= { Y 1 ,y 2 ,…,y n }, the calculation formula of the entropy of Y is as follows: Wherein p (y i ) represents the probability of occurrence of y i ; The conditional entropy calculation formula is as follows: Wherein H (Y|X) represents the entropy of Y under the condition that X occurs, p (X, Y) represents the probability that X and Y occur simultaneously, and p (y|X) represents the probability that Y occurs under the condition that X occurs; the information gain is calculated as follows: , The information gain ratio calculation formula is as follows: ; and dividing the result by calculating two parts of a numerator and a denominator respectively to obtain the value of the information gain rate.
5. The visual detection-based power distribution tower pole deformation monitoring system according to claim 2, wherein the establishment of the influence factor model is specifically that a neuron performs weighted summation on a plurality of received inputs, a result obtained by the weighted summation is compared with a threshold value, and finally, an output of the model is obtained by activating a function; The mathematical formula can be expressed as: Where y k denotes the output result, f (x) denotes the activation function, x i denotes the factor of the input, w ki denotes the weight, and b k denotes the bias.
6. The visual detection-based power distribution tower pole deformation monitoring system according to claim 1, wherein the meteorological disaster monitoring specifically refers to dividing time series precipitation data into a limited hidden oscillation mode and a trend component based on integrated empirical mode decomposition of wavelet analysis, and using the integrated empirical mode decomposition-Markov model for predicting precipitation by using sparse samples.
7. The visual detection-based power distribution tower pole deformation monitoring system according to claim 1, wherein the landslide hazard monitoring specifically refers to real-time target detection and positioning by converting a target detection task into a regression problem through YOLOV algorithm.
8. The visual inspection-based power distribution tower pole deformation monitoring system of claim 1, wherein the access of the monitoring equipment comprises cloud edge end integration, equipment templates, equipment examples and multi-protocol support.
9. The visual inspection-based power distribution pole deformation monitoring system of claim 1, wherein the edge algorithm model establishment comprises model management, model release and model update.
10. The visual detection-based power distribution tower pole deformation monitoring system of claim 1, wherein the Beidou data processing comprises monitoring data acquisition and preprocessing, compression algorithm self-adaptive selection and Beidou short message transmission.

Description

Power distribution tower pole deformation monitoring system based on visual detection Technical Field The invention relates to the field of power transmission and transformation monitoring, in particular to a power distribution tower pole deformation monitoring system based on visual detection. Background The problems of data acquisition difficulty, transmission difficulty, analysis difficulty and the like are more remarkable due to the large number of towers and wide distribution, the towers and the monitoring of the environments where the towers are located, the additional value of the towers serving as the infrastructure cannot be fully utilized, particularly in areas with a large number of mountains, when natural causes such as earthquakes, rainfall and snow melting or ergonomic activities such as excavation of slope feet, pile loading on the upper part of a slope body, blasting and the like occur, natural disasters such as torrents, landslide and debris flow can occur, and the torrent disasters are caused by short-time heavy rainfall, so that multiple people are in distress, the towers are damaged, if one-step discovery can be carried out early, one-step emergency can be carried out early, and the loss can be reduced. However, the existing monitoring mode of the power distribution tower pole only realizes collection of the health data and the image data of the tower by deploying intelligent terminals, cameras and other sensor equipment on the tower, but is not provided with monitoring on disaster conditions such as debris flow, landslide, flood and the like, and the early warning capability of the tower can be reduced by a single monitoring mode. Disclosure of Invention The application provides a power distribution tower pole deformation monitoring system based on visual detection, which has the advantages that the deformation condition of a pole is pre-warned through monitoring the surrounding environment of the pole, a monitoring algorithm model and a data and image compression algorithm model are deployed in an intelligent terminal to be used for realizing the analysis and compression of data and images, and the data and the image compression algorithm model is uploaded to a monitoring center through a Beidou module to be used for realizing the monitoring and pre-warning of a remote pole and the surrounding environment of the remote pole, so that the technical problem provided by the background technology is solved. In order to achieve the aim, the application adopts the following technical scheme that the power distribution tower pole deformation monitoring system based on visual detection comprises monitoring of debris flow susceptibility, monitoring of weather disasters, monitoring of landslide disasters, access of monitoring equipment, establishment of an edge algorithm model and processing of Beidou data; The monitoring of the debris flow susceptibility is specifically to evaluate the debris flow susceptibility by adopting a machine learning algorithm, and the monitoring of the debris flow susceptibility is divided into four processes, namely, pretreatment of data, construction of a susceptibility evaluation model, partitioning of the debris flow susceptibility according to model results and evaluation of model performance; The monitoring of the meteorological disaster specifically refers to the study of annual change according to precipitation data; the monitoring of landslide disasters specifically refers to training a model for identifying landslide by using a network of YOLO V5 in deep learning; The access of the monitoring equipment is specifically based on a cloud edge end integrated architecture, and millions of monitoring equipment access capacity, monitoring data visual analysis and early warning capacity and remote operation and maintenance management capacity are built; The establishment of the edge algorithm model specifically refers to that a manager remotely updates the algorithm model according to monitoring requirements of different regions and different periods, and research and development of the monitoring center algorithm model and training results are applied to the edge side; the Beidou data processing specifically refers to the fact that the Beidou data processing method is used for comprehensively improving the efficiency of data transmission of monitoring data by satellite communication through a data compression algorithm. . Preferably, the monitoring of the susceptibility of the debris flow comprises the selection and classification of the influence factors of the debris flow, the analysis of the importance of the influence factors and the establishment of the influence factor model. Preferably, the selection and classification of the debris flow influencing factors adopts a natural discontinuous method to determine an optimal classification interval, and the concrete operation is that 11 influencing factors are selected, wherein the 11 influencing factors respectivel