乘性噪声图像的偏微分方程复原模型  

A Partial Differential Equation Image Restoration Model for the Image with Multiplicative Noise

作　　者：时文俊[1,2] 周丽雅[2] 郭从洲[2] 秦志远[2] 

机构地区：[1]郑州升达经贸管理学院,河南郑州451191 [2]信息工程大学,河南郑州450001

出　　处：《测绘科学技术学报》2015年第6期594-599,共6页Journal of Geomatics Science and Technology

基　　金：河南省基础与前沿技术研究计划项目(132300410097)

摘　　要：针对含有乘性噪声退化图像的复原问题,提出了一种自适应的、良态的偏微分方程复原模型。该模型是在对该类退化图像统计学规律分析的基础上,通过对正则项的自适应修正和保真项中模糊算子的改进构造的。测试实验数据表明,复原后的图像峰值信噪比与AA模型相比提高了5 d B左右,边缘保持指数比AA模型提高了1倍左右。该模型在对含有服从Gamma分布的乘性噪声和高斯模糊的退化图像复原时,能够在去除噪声的同时保持边缘,复原能力优于其他常见的模型。In order to restore the degraded image with multiplicative noise, an adaptive partial differential equation restoration model was studied and proposed in this paper. This model is based on the class of the degraded image statistical analysis of the law, the adaptive correction of fidelity term and the improvements in the fuzzy operator structure of regularization term. From the experimental data, the peak signal to noise ratiovalue improves about 5dB and the edge preservation index value increased nearly twice as much as the AA model. During restoring the degraded image with the multiplicative noise that follow Gamma distribution and Gaussian blur, this model can keep the edge while remove the noise, the capability of restoration of this model is better than others model.

关 键 词：图像复原 自适应 偏微分方程 峰值信噪比 边缘保持指数 

分 类 号：P237[天文地球—摄影测量与遥感] TP391.9[天文地球—测绘科学与技术]