基于噪声建模的遥感卫星图像分层去噪算法  

Remote Sensing Satellite Image De-noising Based on Noise-modeling

作　　者：唐中奇[1] 孙继银[1] 郭莉莎[1] 梅生[1] 

机构地区：[1]第二炮兵工程学院计算机与指挥自动化系,陕西西安710025

出　　处：《微电子学与计算机》2012年第2期31-35,共5页Microelectronics & Computer

基　　金：国防十一五科研项目

摘　　要：在遥感卫星成像机理分析的基础上,提出一种分层去噪算法,该算法引入了噪声分类建模的思想,根据噪声模型的特殊性分层降噪,最后进行轮廓填充恢复.实验结果表明,该算法不仅能有效抑制多类噪声,也能很好地保持图像的边缘细节信息,在经典去噪评价指标信噪比SNR、均方误差MSE上也都达到了很好的实验标准,具有良好的去噪性能.This paper promotes a new method to de-noise at different stages.Basing at the principle of remote sensing devices to capture images,it introduces an idea of noise-modeling,then de-noises images by the characteristics of different noise models,and finally,recovers the edge and details of the images.Experiments prove that,it can not only de-noise multiple noises efficiently,but also keep the edge and detail information well.It also tests well in classical targets such as SNR and MSE.In a word,this method functions well in de-noising.

关 键 词：遥感卫星图像去噪 噪声建模 分层去噪算法 轮廓填充恢复 

分 类 号：TP751.1[自动化与计算机技术—检测技术与自动化装置]