Bistable stochastic resonance mechanism and its application to the image restoration
双稳态随机共振机制及其在图像复原中的应用

The restoration of the image with strong noise has been playing a key role in improving the visual effects of the image, furthermore, enhancing the performance of the image analysis or processing. This paper studied the aperiodic response of the stochastic resonance in the bistable nonlinear system. Then, by using of the multi-directional Hilbert scanning method, the two-dimensional gray image was converted to several one-dimensional and aperiodic signal sequences, which maintained the space correlation of the image pixels. Finally, the restoration of the gray image with strong noise was achieved by using the stochastic resonance synergy of the bistable system image signals and noise. The experimental results showed that the image restoration method based on the stochastic resonance could effectively suppress the noise and make the image recurrent well in detail. Especially in the low SNR cases, this method was obviously superior to the traditional methods of image restoration in the aspects of the SNR improvement and gray hierarchy.