Research On Image Denoising Based On Improved Wavelet Threshold

Image information plays an increasingly important role in life and production.However,it is difficult to avoid noise generated during image acquisition and transmission,which makes image denoising a crucial task.There are many methods for image denoising.The filtering method is a relatively original denoising method.This method has a narrow surface and a general effect.Recently,the research frequency is relatively high is the wavelet threshold denoising method,the initial use of soft threshold function or hard threshold function to denoise,but both have large defects,hard threshold function is not continuous,soft threshold function Denoising the image is easy to blur the image.Some researchers have proposed some improved threshold functions,but they are more or less flawed.Either reduce the deviation and the threshold function is discontinuous,or the threshold function is continuous but the deviation is slightly larger,so there is still much room for optimization of the threshold function.In order to improve the multiple threshold functions proposed by previous researchers,a new threshold function is proposed.This function has the following advantages: first,the improved threshold function is a continuous function,so that it can avoid the pseudo Gibbs effect after image denoising;second,two adaptive adjustment parameters are introduced in the improved formula,The adaptive adjustment parameters can not only make the improved threshold formula be between the soft and hard threshold formulas,but also can integrate the advantages of the advanced threshold formula proposed by several other researchers,making it more flexible The advantages of hard functions and other improved threshold functions adjust different parameters as the number of wavelet decomposition layers increases.The improved new wavelet threshold function can better solve the problem of excessive deviation after image denoising.By comparing with previous methods,it is found that the new and improved wavelet threshold method has better effect on image denoising.In order to verify the superiority of the proposed wavelet threshold function,multiple sets of simulation experiments were carried out using MATLAB,and the noisy images of two plants collected in real life were used as experimental objects,and it was combined with median filtering,mean filtering,hard threshold,Soft threshold,half threshold,and several thresholds improved by researchers are compared.Calculate the signal-to-noise ratio and peak signal-to-noise ratio,and compare the experimental results and find that the newly improved wavelet threshold function is smaller than the other methods,and the peak signal-to-noise ratio is larger.The new improved threshold method is better than the above-mentioned method and the effect Better.