Self-adaptive wavelet threshold denoisingbased on multi-resolution analysis
COMPUTER MODELLING & NEW TECHNOLOGIES 2014 18(12B) 524-528
School of Computer Engineering,Jiangsu University of Technology,Changzhou 213001,Jiangsu,China
Various noises are usually mixed in the collection, processing or transmission of digital image, which reduces the image quality and which is bad for the subsequent image analysis; therefore, the image denoising processing is an essential link to conduct subsequent image analysis. With continuous development and improvements of wavelet theory, its excellent time-frequency characteristics have led to its extensive applications in image denoising. By analyzing the basic principle of wavelet threshold denoising, this paper has proposed a denoising algorithm of self-adaptive wavelet threshold. This algorithm decomposes and reconstructs the signal by using multi-resolution analysis; designs and constructs appropriate threshold function and realizes a new self-adaptive threshold denoising algorithm by optimizing the threshold with the threshold function. The experimental result demonstrates that compared with median filtering algorithm and mean filtering algorithm, the algorithm of this paper can improve the signal to noise ratio; maintain the detail information and texture features of the image over denoising and have better denoising effects.