基于卡尔曼滤波的光学遥感影像高精度复原处理  

High Precision Image Restoration Based on Kalman Filter for Optical Remote Sensed Images

作　　者：张炳先[1] 王密[2] 潘俊[2] 

机构地区：[1]北京空间机电研究所,北京100076 [2]武汉大学测绘遥感信息工程国家重点实验室,湖北武汉430000

出　　处：《武汉大学学报（信息科学版）》2015年第7期964-970,共7页Geomatics and Information Science of Wuhan University

基　　金：国家973计划资助项目(2014CB744201;2012CB719901);国家863计划资助项目(2011AA120203);长江学者与创新团队发展计划资助项目(IRT1278);国家自然科学基金资助项目(41371430);全国博士学位论文作者专项资金资助项目(201249)~~

摘　　要：光学卫星成像系统调制传递函数(MTF)的准确量测是高质量影像复原的基础。传统的MTF测量方法忽略了卫星平台振动、影像噪声等因素的干扰,导致测量结果与真实值存在较大偏差,不利于影像质量的提升。本文在分析现有MTF测量方法的基础上,提出了基于卡尔曼滤波的高精度MTF测量方法,该方法利用卡尔曼滤波对实测的线扩展函数(LSF)进行迭代处理,获取无干扰的LSF,为影像复原质量的提高奠定基础。本文利用国产高分辨率卫星成像数据进行实验,采用边缘能量、对比度、奈奎斯特频率值作为复原前后影像质量评价的依据,实验结果表明,采用本文方法获取的MTF进行复原的影像无论是在边缘保持还是噪声抑制方面都优于传统方法。High precision MTF measurement is the basis of high quality image restoration.Given the presence of noise in images and vibration from the payload,traditional MTF measurement based on the target image will produce a biased result,and the biased result will introduce new noise after image restoration.In this paper,based on analysis of characteristics and limitations of traditional image restoration method,we propose an image restoration approach based on high precision MTF measurement using the Kalman filter.This approach firstly uses Guassian fitting to obtain theoretical value of the line spread function from the measured value,then it uses the Kalman filter to obtain the true value of the line spread function from theoretical value and measured value.Experiments on TDI-CCD images show that the approach proposed in this paper yield better performance than traditional image restoration,especially for the water areas which contain less texture and city areas which contains rich texture.

关 键 词：光学成像系统 影像复原处理 卡尔曼滤波 调制传递函数 高斯拟合 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]