Research And Application Of Image Proceesing Technology For Damage Inspection System Of Overhead Powerline Porcelain Bottles

The maintenance of the high voltage lines depends on the workers mainly in our country now, which has some disadvantages such as high labor intensity, high costs and high risk. Therefor developing new a new kind of equipments instead of workers to realize automatioc inspection on high voltage power lines has become a hot research topic recently. The self-walking patrol robot for high voltage power lines which integrates multiple inspection technologies can realize visual surveillance. The invention of the patrol robot breaks out a completely new concept and method for the high voltage lines maintenance. The research for vision inspection system in patrol robot under 863 project sponsored is presented in the paper.The recognition and damage inspection of porcelain bottles in the vision system of patrol robot is mainly studied in the paper. The damage inspection of porcelain bottles is implemented based on image processing technology. The simulation is completed by Matlab and the inspection system is designed at last. The contents are as follows:Firstly, the image of porcelain bottles is dealt with the image preprocessing technologies. After preprocessing, the features are extracted from regions. Preprocessing technologies such as grey scale transformation, de-noising, segmentation, morphologic processing, region labeling and edge tracking are used. Comparative analysis is made in the paper among fixed threshold, iterative method and Otsu algorithm. Through the analysis of porcelain bottles images, characteristic regions in the center of porcelain bottles are taken as the basis of location. The features are extracted based on the special shape of this characteristic region. According to the rotation, scale and translation invariance, features including circularity, slenderness, duty cycle, aspect ratio and Tchebichef moment invariants are selected in the paper.Secondly, the porcelain bottles are located based on pattern recognition technique in image processing. Recognition of characteristic region is completed by probabilistic neural. In order to improve the accuracy rate, genetic algorithm is used to optimize the smoothing parameter and hidden centers vectors. At last, the location is completed according to the ratio between rectangular frame of characteristic region and length to width of porcelain bottles.Thirdly, the research on damage inspection algorithm of porcelain bottles, including surface damage and edge notch inspection, is implemented. Edge detection is used in the surface damage inspection region. The damage of porcelain bottles is judged by the inspection of length and gray value of isolated narrow chains and the gray level rate of longitudinal tangent.Finally, the damage inspection system of overhead power line porcelain bottles in the overhead power lines is designed. The simulation of patrol robot's visual system is accomplished by hardware configurations such as CCD (Charge Coupled Device), rotational station and image acquisition card. Da Heng image development library and Visual C++6.0 are used to accomplish the software design and realize the collection and processing of porcelain bottles image.