Comparison Of Non-Bayesian Wavelet Filtering Method And Research Of Threshold Value Filtering Algorithm

The appearance of wavelet has aroused the widespread interest in the field of mathematics, its theory and application obtained the unprecedented development, and it has become a main research direction of the signal and image processing. The wavelet theory and application as an important branch, the wavelet filter's theory and application also obtained the prodigious development for several years, which symbolizes the emergence of a kind of new signal filtering method.The wavelet filter's mechanism is based on the different property of wavelet coefficient's scale between signal and noise, it uses the corresponding rule and it processes the signal containing noise by non-linear method that contains extraction, choice, cutting the signal wavelet coefficients and so on, in order to eliminate the noise. The research of wavelet filtering mainly has three directions, including the mold maximum value restructuring filtering based on signal irregularity, the space correlation filtering based on the relativity between signal scales and the wavelet threshold value filtering based on the decorrelation of wavelet transformation, and they obtain a more successful application in medicine image filtering, SAR image filtering and compression, and other fields.The method of wavelet threshold value filtering is a kind of the simplest realization, the smallest calculation, so it has been widely applied. But its threshold value's selection is quite difficult, although Donoho has proved theoretically and has found the general threshold value, the results are unsatisfactory in the practical application. Therefore, the primary content of research is to analyze and compare each non-Bayesian filtering method and the improvement threshold value method in the field of wavelet in this paper. And that there are two aspects in the improvement of threshold value method: First, it proposes the GID threshold value method based on wavelet transformation and grey incidence degree, and it will be applied to the signal filtering and the image compression; Second, it mainly research on the parameter selection algorithm of semi-soft threshold function in the wavelet threshold value function. Combined with the grey incidence degree theory and fuzzy theory, a new parameters' algorithm will be proposed, and it can reduce the complexity of algorithm, finally, it will be applied to the different signals, and the results are satisfactory.In this paper, the GID threshold value method as well as the improvement of semi-soft threshold value function parameter has obtained famous effect by simulating the different property signal and image by means of MATLAB. Their in-depth research and application will play a positive role in the wavelet transformation with the processing of the signal and image.