Image Enhancement Method Based On Aperiodic Stochastic Resonance And Its Application

The generation condition of stochastic resonance requires the interaction of noise and nonlinear systems.The nonlinearity of the system plays a decisive role in the stochastic resonance effect.The nonlinearity of the potential function can be characterized by the concave or convex of function and number of function extremums.In this paper,in order to achieve image denoising and low-illumination image enhancement,the characteristics of image stochastic resonance under the action of different potential functions are studied.The relationship between stochastic resonance effect and image enhancement is analyzed,and a low-illumination image enhancement method based stochastic resonance is proposed.It has good application prospects in night license plate recognition.The main work is as follows:(1)A parabolic bistable potential function is constructed based on the classical bistable potential function.By analyzing the influence of the concave or convex of bistable potential function inner and outer sides on the Kramers escape rate,the nonlinear system output approximate analytical expression for the peak signal-to-noise ratio is established under the action of images and noise,and the relationship between the concave potential function and the resonance effect is elucidated.Theoretical analysis and numerical simulation results show that by adjusting the concaveness of the potential function,the stochastic resonance effect can be effectively changed,so that the resonance effect is the strongest and the image enhancement effect is the best.(2)A quad-stable potential function with four minimum points and three maximum points is constructed.The approximate analytical expression for the image variance with respect to the noise intensity is derived.The characteristics of quadstable images stochastic resonance are analyzed and a low-illumination color image enhancement method based on random resonance is proposed.The results show that the four-stability system has a higher utilization rate of noise and has better image denoising and enhancement effects.(3)Combining subjective and objective evaluation criteria,image denoising and enhancement experiments are performed on common images such as people,landscapes,buildings and highway.The advantages of image enhancement method based on stochastic resonance are discussed,and the method is applied to night license plate detection.Experimental results verify the effectiveness and feasibility of the proposed method in image denoising and low-illumination image enhancement.The enhancement of stochastic resonance effect can further improve the quality of images,and has a good application prospect in night license plate detection.