In-Service Insulator Image Denoising

A Poisson-Gaussian (PG) denoising method for inservice insulator image is proposed. The existing spatial-based method can avoid the preserve more detail information than the transform-based method. Various transforms have been used for image denoising. However, the transforms have disadvantages such as aliasing and lack of directional selectivity. Meanwhile, the conventional redundant transform-based image denoising approaches require more memory and consume more time. In order to capture detail information of images more efficiently, and reduce the computation cost, we propose a hybrid method to improve quality of denoised insulator image in dual domain. Experimental results show that the proposed method is able to significantly improve the PG denoising performance in visual inspection.