The Method Of Image Detection For Broken Strands Of Transmission Line

With the promotion and application of inspection technology of transmission line by helicopter and the construction of smart power grids, the intelligent detection of conduc-tor defect in the transmission line has been a hot issue. The steel core al stranding wire, which is the main material for China’s long-distance electricity transmission and distribu-tion, may lead to the local damage and eventually broken strand with the influence of the environment, such as wind power, icing and temperature. The research of automatic de-tection for Broken strands of Transmission line has a very important practical signific-ance. This paper focuses on how to detect the wire broken stocks defect in the transmis-sion line efficiently and accurately with the image processing and recognition technolo-gy.According to the characteristic of the image transmission line, such as complex background, contains more random noise and so on, an image pretreatment process on power line image is proposed. This step mainly includes transmission line image en-hancement processing based on the piecewise linear gray level transformation, image denoising processing based on the transfinite pixel smoothing method and transmission line edge detection processing based on the improved Sobel operator. This step effec-tively enhances the contrast of the transmission lines in the image, eliminate random noise, and make the transmission line image having a better effect of the edge detection. Then compare and analyze several common line detection methods, and an algorithm is proposed to extract straight line based on the combination of random Hough transform and regional growth of gradient direction information. This method solves the problems of the random Hough transform effectively and improves the efficiency and accuracy of the line extraction. In the meantime, the transmission line can be identified using the two prior conditions that the pixels gray level of the transmission line is low and the trans-mission wire is a long line. Finally, this paper proposes a method to detect and identify the conductor defect in transmission line based on target pattern matching and provides an intersection structure template of the broken stock point. Experimental results shows that the method proposed can detect the broken strands defects of the transmission line better.