Inspection Of Damper During Power Line Tour By Helicopter

With the rapid development of the transmission network’s construction in China,the coverage of power supply network has accounted for more than90%of ourcountry area, and there are many difficulties on the safety inspection in cold, high-altitude transmission line. Routing inspection transmission lines by using helicopterscan solve these problems, at the same time the accuracy and the speed of inspectionshould be increased. This paper will deeply analyze and study the method of imagerecognition on transmission lines damper.First of all, preprocessing the image collected by helicopters, including imagefiltering, image segmentation and fuzzy recovery. Combining with the method ofaverage filtering and median filtering to filter out Gaussian noise and impulse noiseon the images then evaluate experiments. Using the method of Threshold-basedimage segmentation to highlight the interested parts on the image then evaluateexperiments. Since the images are collected in a helicopter with high-speedmovement, they are prone to motion blur, this paper will analyze the restoration offuzzy images, and evaluate experiments by using the method of restoring the fuzzyimages based on gradient domain.Secondly, this paper focus on the technique of damper edge detection aimed toclear and blurred images. For the clear images, I designed Hessian matrix and ridgeline tracing combining method to do the edge detection on anti-vibration hammerimages, compared by experiments which use the classic canny method, Hessianmatrix method. The method on this paper has the advantages of integrated detectionon accuracy and speed, therefore this is more suitable for a helicopter inspection oftransmission lines. Aiming at blurred images, this paper designs an improved balloonforce Snake model to do the edge detection with noisy blurred image, makes the experimental comparison,extraction and the evaluation to the mean square on actualedge position,and contrasted to enhanced balloon force Snake model method andHessian matrix binding method to assay results. Experimental results show that theimproved balloon force Snake model method has more ability of noise and blur toadapt.