Research On Insulators Segmentation And Location For Aerial Image Based On PCNN

The method of Patrol inspection with helicopter, which has considerable advantages:efficiency, rapidness, reliability, low-cost, disregard of distance etc, is an important way that inspection of transmission lines now." Helicopter-based inspection, supplemented by manual inspection" is development of high voltage and EHV line inspection. Insulators images acquired in patrol inspection provide basis data for intelligent fault diagnosis. Insulator, as an important part of a transmission line, that is operated in the field long times,is vulnerable to be affected by operating environment.Broken insulator is a serious threat to the normal operation of the transmission line.Image processing technology and GPS technology is applicated in the field of the electric inspection,it is possible for helicopter automatic inspection. As image segment is an important technology in image processing, it plays an important role in extraction of insulator and detection of late diagnosis. Aim at insulators image that have features like complicated background,more noise pollution and highter brightness of insulators,we apply the new generation artificial neural networks----Pulse Coupled Neural Networks (PCNN) to insulators image in helicopter patrol inspection. Compare to the traditional artificial neural network, PCNN has biological characteristics, it does not need accurate training and has good characteristics of pulse emission,parameter controllability.So it is widely used in image processing and pattern identification.The paper used PCNN to extract the insulators of aerial insulators image.To aerial insulators image,we used the method of image segmentation to divide the image into different area.According to the set threshold value,when the gray of the target area (insulators) is greater than or equal to the set threshold value,it will be pulsing,pulsing pixels transmission connections with neighboring pixels;But below the set threshold value,it will be suppressed without lgnition,eventually the pixels of the whole target area is pulsing,then extract the insulators from the image effectively.At last, the paper compared the segmentation results that processed by PCNN on the original color images with the gray images,so it pointed out that the segmentation results of the B-channel in RGB channel of the color insulator images.And compared the images that processed by PCNN with common segmentation algorithms,such as Bimodal histogram algorithms, Otsu and so on. From the results, it pointed out that the PCNN model showed expected results in processing the insulators images.Then on the segmented images, insulators is identificated and located in original images by Generalized Hough transform.