Image Denoising Algorithm Based on Information Feature Subspace Uniform Traversal

Remote image acquisition in strong interference environment usually contain a lot of noise, the image quality is not good, the need for image denoising, image recognition and imaging performance improvement. The traditional image denoising methods using wavelet transform noise smoothing algorithm, the emergence of visual offset ratio is small in the case of signal to noise, image denoising effect is not good. This paper proposed an image de-noising algorithm information feature subspace based on uniform traversal. Based on triangular mesh texture partition structure model, the establishment of the noisy image linear model of direction relations by linear mode between approximately orthogonal design implementation, image filtering, noise isolation, with global non obvious mutation information extraction methods, they are ergodic addressing way based on the characteristics of image noise filtering to realize the traverse, implementation improved image denoising algorithm. The simulation results show that the algorithm has better performance in image denoising, it can improve on the quality of imaging and identification probability fuzzy image with noise, peak signal to noise is relatively high, and it has application value in the fields of remote fuzzy image recognition.