摘要
传统的图像滤波复原技术都是把噪声当成一种有害的干扰加以消除,但随着噪声的强度增加,传统的图像复原办法对强噪声背景下的图像复原效果很差。本文主要基于Hodgkin-Huxley(H-H)神经元阈上非周期随机共振原理,通过一种自适应调节的方法不断添加噪声实现图像的随机共振,从而达到最佳的图像复原效果。实验结果证明,相对于传统的图像复原方法,本文所提出的方法在强噪声的背景下对图像的恢复有更好的效果,随着噪声强度的变化具有比传统方法更好的鲁棒性,为图像复原提供了一种新思路。
The traditional image filtering restoration technology regards the noise as a harmful disturbance and ehminates it,but if the noise intensity increases, the traditional image restoration means is very bad to restore the image under the strong noise background. According to the principle of stochastically resonating the Hodgkin-Huxley (H-H) neuron thresh- old, the palJer increases the noise realization image unceasingly through an auto-adapted adjustment method of stochastic res- onating, thus achieves the best image restoration effect. The experimental results prove that, compared with the traditional image restoration method, this method has a better effect in image restoration under the strong noise background,and it pro- vides a new mentality in the image restoration, and along with the noise intensity changes this method has better robustness than the traditional one.
出处
《计算机工程与科学》
CSCD
北大核心
2009年第5期46-48,共3页
Computer Engineering & Science
基金
国家自然科学基金资助项目(60374047)
关键词
图像复原
H-H神经元模型
阈上非周期随机共振
自适应控制
Otsu图像分割
峰值信噪比
image restoration
Hodgkin-Huxley model
self-adaptive control
suprathreshold aperiodic stochastic resonance
Otsu image segmentation
power signal-to-noise ratio