Intelligent Restoration Algorithm to Line Similar Pixels in Noisy Image

Abstract

For restoration line similar pixels from noisy image, the disadvantages of some current image processing methods are analyzed. Then a new directional operator set with standard uniform is designed to detect the anisotropy of line pixels. Learned from manmade game schedule of washing out and sporting, a non-orthogonal and non-PR(perfect reconstruction) filtering approach is proposed. With the program of measuring, washing out, competition and tight support domain filtering, simplified by the plot of from simple to complex, straight line estimation to zigzag fitting, relative algorithm is implemented in an orderly manner. The effectiveness and consistency of proposal algorithm are checked at different noise level, and the satisfactory lines restoration results are obtained.