Image Denoising Method Based on Hybrid Deep Dictionary Learning and Elastic Threshold

Abstract

The application of machine vision in the modern industry makes intelligent manufacturing possible. At present, circuit board related fault detection, such as broken wires, missing solder joints, and other problems frequently occur, the manual recognition speed is slow and the false detection rate is high. In response to this problem, this research proposes an image noise reduction processing method that combines deep dictionary learning and elastic thresholds, which can serve computer image processing, thereby greatly improving the detection speed of printed circuit boards, and ultimately serving the processing and production of modern electronic products.