Recognition And Research Of Anomaly Map Of Transmission Line Inspection Based On Computer Vision

With the continuous development of the computer vision theory and gradually mature, avariety of application technology based on computer vision are constantly emerging, they playimportant roles in production and life. This paper puts forward some methods to solve theproblem of transmission line fault detection with the actual problems encountered in theproduction based on the computer vision technology, in order to achieve the purpose oftraditional technology innovation, and further to deep understanding of the computer visiontechnology.Regular inspection and maintenance of the transmission line is a high difficulty, high risk,heavy workload task. The transmission line widely and complicated location, the power lineand tower accessories will be affected by many uncertain factors. If not can be timelytreatment and rehabilitation, it will lead to serious accidents, to bring great hidden danger forthe power transmission. Therefore, regular inspection and maintenance of overheadtransmission lines is an important measure to maintain the steady development of the nationalelectric power. For a long time, most of operation maintenance transmission lines workmainly rely on manual operation, the inspection accuracy is low, the labor intensity is big, andexisting inspection blind area, forest and wild animal diseases also poses security risk to theinspectors. With the continuous development of the key technology of machine vision andimage processing technology, using patrol helicopters power inspect line instead of manualoperation. The helicopter patrol inspecting transmission line, with the big, wide viewing angle,high efficiency and other advantages. Anomaly map of transmission line inspection based oncomputer vision recognition technology is mainly set important parts of the transmissionline in the image information through the helicopter patrol inspection image, analysisImage anomaly characteristics of key frame image, to get defect recognition of transmissionline of the key image information. We use the method of image processing and computergraphics theory related to OpenCV, aiming at the abnormal inspection figure three commonsituations--transmission lines, transmission lines and bifurcation suspension insulator,presents a better method to identify. The method proposed in this paper can improve the inspection efficiency of abnormal pattern recognition, and has a certain practicality.For transmission line forking, we firstly do the basic process on aerial image, includinggray processing, thresholding, edge detection and morphological closing operation, thenextract transmission lines using the Hough line detection, finally analyze slopes oftransmission lines combined with actual errors to judge the forking case.For transmission line hanging spoils, after the basic process on the aerial image, weextract normal transmission lines by Hough line detection, then make the convolutionoperations at small areas around the transmission lines, finally calculate rate of "dirty spot" ofthese areas combined with actual errors to determine whether transmission line hangs spoilsor not.For insulator lost, we extract the H component image after converting orginal image toHSV space, then correct the orientation of the image, and use H component image of normalinsulator image as a slice to match contours, what’s more, we filter wrong matched regions(including the vertical and horizontal direction), according to the symmetry and locationfeatures of insulators, ultimately determine whether the insulator is lost or not based on theircharacteristics.Recognition technology of anomaly map of transmission line inspection based oncomputer vision solved the inconvenience of manual testing, the advantages of artificial isirreplaceable. Of course, there still exist some shortcomings, we will gradually be corrected inpractical application.