Study on Image Noise Reduction Algorithm at Improved NL Means Based on Color Information

Abstract

For the random noise of digital image processing is embedded in the old films, to study the improved Non Means algorithm in Local. Through the whole area of the surrounding pixels of gray distribution in contrast to the overall, according to the gray level distribution similarity obtained decision weights, to study and calculate the corresponding each pixel set color Information adaptive parameters, and then calculate the similarity, screened based on the template. In order to achieve the goal of eliminating the noise by modifying the value of the target pixel region, and using the weighted average value instead of the target pixel region value. Finally, the feasibility and the scientific nature of the noise reduction performance are simulated by referring to the image noise elimination precision and image quality parameters.