基于区域分割的混合PDE模型遥感图像去噪  

A PDE Based Hybrid Model for Denoising Remote Sensing Image Based on Regional Segmentation

作　　者：张帆[1] 葛玉荣[1] 

机构地区：[1]中国海洋大学信息科学与工程学院,山东青岛266100

出　　处：《计算机测量与控制》2013年第11期3090-3092,共3页Computer Measurement &Control

基　　金：国家自然科学基金(40976060)

摘　　要：针对遥感图像所含噪声为CCD噪声,结合纯各向异性扩散模型和四阶PDE模型在图像去噪中各自的优势,提出了一种基于区域分割的混合PDE模型遥感图像去噪算法;实验表明,该算法能够在有效滤除噪声的同时,很好地保留图像的边缘和细节特征,峰值信噪比达到了25.5649dB,清晰度达到了10.8265。According to the characteristic that remote sensing image mainly contains CCD noise, and taking advantage of pure anisotropic diffusion model and 4--order PDE model in image denoising, this paper proposes a method a PDE based hybrid model for denoising remote sensing image based on regional segmentation. Experimental results illustrate the proposed algorithm can not only efficiently remove noise, but also simultaneously retain image edge and details characteristics, Peak signal -- to -- noise ratio is 25. 5649dB, the resolution is 10. 8265.

关 键 词：遥感图像 区域分割 偏微分方程 去噪 

分 类 号：TP303[自动化与计算机技术—计算机系统结构]