PD image recognition using fractal features and statistical parameters

Starting with partial discharge (PD) artificial insulation defects designation and HV defect model tests, a suitable set of PD pattern recognition features of PD images, consisted of fractal features and statistical parameters, are determined and then used as input variables to a back-propagation neural network for the purpose of classifying the PD patterns. In this procedure fractal dimensions and the 2nd generalized dimensions of original PD images and fractal dimensions of high gray intensity PD images are proposed and computed by MDBC method, and thereafter moments and correlative statistical parameters are studied for recognition of PD images. Following the illumination of the basic mathematical concepts regarding the above parameters, final recognition results for experiment PD data samples show good performance of the proposed method which appears promising for future work.