A condition monitoring system based on dyadic wavelet transform using thermal image

Abstract

Wavelet analysis is broadly and effectively utilized in image processing and analysis. Because of the simple and fast algorithm, the dyadic wavelet transform has generally used. A dyadic wavelet transform gives an excellent output because of the different levels of the wavelet coefficient of the image. The proposed method presents an optimal value such as mean, standard deviation and entropy for the decomposition and reconstruction of the thermal image. The original results demonstrate which technique can find the induction motor faults clearly and effectively.