Application of the sparse decomposition algorithm in the film defect denoising

This paper aims to extract the exact defect characteristics of the thin film surface of the lithium battery by sparse decomposition algorithm. An appropriate atomic function was selected and the sparse decomposition iteration was conducted on the defect images in the overcomplete dictionary. This value from observation method was taken as the empirical value and applied as the iteration termination condition of the sparse decomposition. Then, the denoised defect images were obtained. The results reveal that the sparse decomposition has a far superior denoising performance to that of the median filtering technique, and can better restore the thin film defects of the lithium battery.