Research On Transmission Line Insulator Object Detection Method Under Complex Meteorological Conditions

At present,intelligent inspection technology has been widely used in the operation and maintenance of transmission lines.After a large number of aerial images are collected by cameras or monitoring sensor equipment,it has become another difficult problem for operation and maintenance inspectors.As a key inspection target,insulators have a wide range of quantities and types.If they are only detected manually,it will not only cost a lot of time,but also can not locate the area in time,which will cause potential safety hazards to the normal operation of transmission lines.Therefore,computer vision technology has become a research hotspot in the field of intelligent inspection of transmission lines.However,the transmission line image is located in the outdoor scene,and there are often complex meteorological conditions,such as haze or rain,which will lead to unclear image and occlusion of insulator target features,which will greatly reduce the quality of the collected image,and then affect the accuracy of insulator detection.This paper focuses on the problem of low image quality of transmission line under complex meteorological conditions.Through the comparative analysis of the current mainstream image dehazing algorithm and image deraining algorithm,an improved algorithm is proposed to improve the image quality under hazy or rainy conditions,and then the target detection of transmission line insulator is carried out to realize the intelligent inspection of transmission line.Firstly,this paper introduces the research status of image processing under hazy and rainy weather conditions at home and abroad in recent years,analyzes the advantages and disadvantages of the existing image dehazing and deraining methods,and designs the overall research scheme.Then,for the insulator image with the above situation,on the one hand,how to solve the problem of dehazing failure of outdoor insulator image in dark channel method in hazy weather is studied.An image dehazing method based on sky region segmentation is proposed,which combines gradient information and edge detection to obtain accurate sky boundary line,so as to obtain accurate atmospheric light value and transmittance and restore the fog free image with good visual effect;On the other hand,the occlusion of rain marks on insulator targets in rainy days is studied,and a image deraining method based on context aggregation detail network is proposed.By separating the high-frequency part of the rainy image,the detail image with rich rain grain information and weak background information is obtained,and the network model is constructed to fully learn the rain characteristics and obtain a clean and clear rain free image.Finally,the insulator images after preprocessing are used to construct the detection integration models under different weather conditions to realize the insulator target detection under complex weather conditions.The results show that the detection accuracy of the proposed dehazing and deraining methods is improved in different models,so as to solve the problem of intelligent inspection of transmission lines under different meteorological conditions.