NONLINEAR IMAGE FILTER FOR REMOVING NOISE

摘要

PROBLEM TO BE SOLVED: To reduce noise without making detailed parts of an image out of focus and affecting temporal frequency by replacing a target pixel in a processing window by weighted average of plural adjacent pixels according to a degree of similarity between the target pixel and the adjacent pixels. SOLUTION: The target pixel in the processing window is replaced by the weighted average of plural adjacent pixels according to the degree of similarity between the target pixel and the adjacent pixel. An absolute value of difference Di between luminance of the target pixel X and luminance of the adjacent pixel Xi is calculated and a differentiation value is generated by a differentiating device 204 in a noise filter 200. The maximum difference Dmax and the minimum difference Dmin which are selected from the absolute values of the difference Di are discovered by a maximum and minimum block 206. A noise level NL which is related to the processing window is calculated by a level calculation block 208. A new pixel value is obtained by an evaluating device 216 of the degree of similarity and a final filter 218 in an evaluation block 214 of the degree of similarity.