Recognition Algorithm of Partial Discharge Pulse Sequence Based on CLAHE Enhancement

Aiming at the diagnosis of the types of insulation defects in GIS, a partial discharge pulse sequence recognition algorithm based on CLAHE enhancement is proposed. First, use the CLAHE algorithm to process the PRPS map to achieve the enhancement of the discharge pulse target, enhance the visual characteristics of the data set, and improve the recognition effect. Then the ULBP of the PRPS map is calculated as the feature vector. Finally, the Adaboost cascade classifier is used to continuously improve the recognition rate until convergence. The experimental results show that the CLAHE enhancement algorithm overcomes the shortcomings of AHE's excessive distortion, and can greatly increase the contrast of the map while retaining sufficient local detail information to achieve the feature enhancement of the data set. Under the same ULBP feature extraction process and Adaboost convergence conditions, CLAHE enhancement increases the upper limit of the recognition rate from 93.36% to 96.09%, expanding the optimization space.