Method for detecting and processing corrosion defect image of tower stay wire on power transmission line

Abstract

The embodiment of the invention provides a tower stay wire corrosion defect image detection processing method on a power transmission line, and the method comprises the steps: collecting a tower stay wire image on the power transmission line through an unmanned plane, segmenting the tower stay wire image through a full convolutional neural network, obtaining a plurality of image blocks, and taking the image block containing a defective tower stay wire as a target image block; extracting a feature vector of the tower stay wire based on the target image block; screening out the feature vectors meeting a preset condition to form a target screening feature set; extracting a typical defect characteristic value of the characteristic vector of each tower stay wire; and performing similarity calculation on the typical defect feature value on each tower stay wire and the target screening feature set to obtain the defect similarity, and judging whether the current tower stay wire has strand breakage or corrosion defects based on the defect similarity, thereby improving the corrosion defect detection precision of the tower stay wire.

Description

technical field [0001] The invention relates to the field of computer technology, in particular to a method for detecting and processing images of corrosion defects of poles, towers and cables on transmission lines. Background technique [0002] At present, with the application of helicopter inspection technology for transmission lines and the gradual advancement of smart grid construction, automatic detection of power line defects has attracted more and more attention. The pole tower cable is the main material of the main tower. Affected by environmental factors such as wind, ice, temperature, etc., it is easy to cause local fatigue damage, and even cause partial broken strands and insufficient tension of the cable. Therefore, the research on the automatic detection of the above-mentioned pole tower cable has great practical significance. [0003] Corrosion of pole and tower cable occurs frequently in long-distance transmission line systems and is very harmful. There are m...

AI Technical Summary

Problems solved by technology

[0004] The identification methods of tower stay wires include traditional feature extraction based on Hough transform, extraction of histogram of oriented gradient (HOG) features of tower stay wires, etc., but these methods have low extraction accuracy due to the single extracted features

In recent years, with the rapid development of artificial intelligence technology, neural network technology has been widely used in the identification of objects. Some people have proposed to identify the tower cable through convolutional neural network, mainly using the input layer + several interleaved convolutional layers and pooling The structural neural network structure of layer + fully connected layer has a single structural mode. By setting a large number of tower cable training sets to train the convolutional neural network, the calculation amount is huge, and once the shape of the tower cable changes, the trained neural network model It will not be able to accurately detect the pole and tower cable, and the migration of the single-mode neural network model is poor

Images