The Research And Application Of Stochastic Resonance In Signal Detection And Image Processing

In communications and other fields, noise exists everywhere and always interferes with the detection and transmission of signals. Unlike the conventional signal processing method to filter out noise, the method of stochastic resonance is rational use of the noise. When it reaches a certain match among the signal, the noise and the nonlinear system, the performance of signal processing can be maximized.This paper mainly studies the phenomenon of stochastic resonance in nonlinear multiple signal detection for additive noise and multiplicative noise and stochastic resonance in the application of image processing based on peak signal-to-noise ratio. The main works are as follows:First this paper discusses stochastic resonance in nonlinear multiple signal detection for additive noise and four representative multiplicative noise(uniform noise, Gaussian noise, Laplace noise, and Cauchy noise) based on the maximum a posterior probability(MAP) criterion in the probability of detection error measure. When the signal is suprathreshold, multiplicative noise and additive noise are always deteriorate signal detection performance. When the signal is subthreshold, additive noise and multiplicative noise can reduce the probability of detection error and improve the performance of signal detection together. The phenomenon of stochastic resonance is generating.Then the paper discusses the phenomenon of stochastic resonance of the multi-subthreshold image which uses the PSNR as characteristic parameter. When the noise is added to the threshold image, the PSNR of the image increase firstly and then reduce with the increasing of noise intensity. The phenomenon of stochastic resonance is generating. A suitable noise can improve the visual effect of the image.Finally the paper discusses the image restoration method based on array bistable stochastic resonance. The restored image not only has more excellent gray layering than the image use of traditional filtering method, but also has more smooth and clear visual perception than restored image based on a single bistable stochastic resonance.