Research On Image Denoising Based On Stochastic Resonance Mechanism Of Saturated Nonlinear System

Italian scientist Benzi et al.Proposed the concept of stochastic resonance,which was used to explain the phenomenon that the glacial and warm climate periods appear alternately in the ancient weather of the earth.The principle is that when the nonlinear system,signal and noise reach a certain coordination relationship,part of the noise energy will be converted into signal energy.Since its introduction,the stochastic resonance mechanism has received extensive attention.Based on the correlation characteristics between one-dimensional signals and two-dimensional signals,this paper extends the research field of stochastic resonance from one-dimensional signals to two-dimensional signals,that is,the field of image processingThe main work and innovations of this article are as follows(1)Based on the saturated nonlinear system,the image is denoised,and the application of the saturated nonlinear system is extended to the field of image denoising.During the experiment,it is necessary to convert the two-dimensional image into a one-dimensional signal.According to the correlation characteristics between the one-dimensional signal and the two-dimensional image,this paper adopts the Hilbert scanning method,the row and column scanning method and the row and column sequential scanning method to reduce the dimension of the two-dimensional image signal(2)A two-dimensional stochastic resonance image denoising algorithm based on a saturated system is proposed.Two-dimensional stochastic resonance is achieved through two one-dimensional stochastic resonances.The static saturation nonlinear system and the dynamic saturation nonlinear system are selected as pre-processors of the image signal.By adjusting the parameters of the nonlinear system,the phenomenon of stochastic resonance is induced,and finally the effect of image denoising is achieved.In the experiment,one-dimensional stochastic resonance and two-dimensional stochastic resonance were respectively performed on the noisy images.The experimental results show that both the one-dimensional stochastic resonance and the two-dimensional stochastic resonance have good image denoising effects,and the two-dimensional stochastic resonance The noise effect is better than the one-dimensional stochastic resonance image denoising effect(3)An array stochastic resonance image denoising algorithm based on a saturated system is proposed.Dynamic saturation nonlinear systems are combined together in an array to enhance the random resonance effect of the dynamic saturation nonlinear system.By adjusting the number of dynamic saturation nonlinear systems in the array and the parameters of the system to achieve a better stochastic resonance image denoising effectThe image denoising algorithms proposed in this paper all use the peak signal-to-noise ratio and the similarity coefficient as the standard to measure the image denoising effect.By observing the effect map before and after stochastic resonance image processing,the peak signal-to-noise ratio state table and the similarity coefficient comparison table,we can see that static saturated nonlinear systems,dynamic saturated nonlinear systems and array saturated nonlinear systems have better noise-containing grayscale images The image denoising effect provides a new idea for the research of stochastic resonance in the field of image denoising.