嫦娥一号卫星CCD立体相机影像超分辨率重建算法  

Super-resolution reconstruction of Chang′e-1 satellite CCD stereo camera images

作　　者：张磊[1,2] 杨建峰[1] 薛彬[1] 梁士通[1,2] 

机构地区：[1]中国科学院西安光学精密机械研究所,陕西西安710119 [2]中国科学院研究生院,北京100049

出　　处：《红外与激光工程》2012年第2期404-408,共5页Infrared and Laser Engineering

基　　金：国家863计划(2010AA122201);国家自然科学基金(60808028)

摘　　要：针对嫦娥一号卫星CCD立体相机空间分辨率不足的问题,运用最大后验概率估计法(MAP)实现了月表影像的超分辨率重建。介绍了嫦娥一号卫星CCD立体相机的成像模型,分析了图像获取过程中的主要影响因素,并建立了相应的超分辨率重建模型。基于该模型,首先采用误差-参数分析法估计嫦娥一号卫星CCD立体相机动态成像光学系统的点扩散函数(PSF);然后将估计的PSF应用到MAP算法所建立的目标函数中,采用共轭梯度法对目标函数进行最值求解;再通过VC软件平台编程实现了对单帧正视月表影像的超分辨率重建;最后从信息熵、清晰度和频谱等方面对重建图像进行评价,结果表明重建图像像质优良。Considering the limited space resolution of the Chang′e-1 satellite′s CCD stereo camera,MAP（maximum a posteriori probability） algorithm is used for lunar images super-resolution reconstruction.Imaging model of the Chang′e-1 satellite CCD stereo camera,analyses chief factors which determine the image quality in brief,and establishes appropriate model for super-resolution reconstruction are introduced.According to this reconstruction model,the error-parameter analysis was used to estimate the point spread function（PSF） of dynamic imaging system of the Chang′e-1 satellite CCD stereo camera firstly.Secondly,the estimated PSF was applied to the objective function established by the MAP algorithm,and the conjugate gradient method（CGM） was used to solve the extreme value of the function.Thirdly,the super-resolution reconstructions of the single frame normal lunar images were prograrmmed on VC software platform.Finally,the reconstructions through entropy,definition and spectrum were evaluated.Results show that the reconstructions quality is of great.

关 键 词：嫦娥一号卫星 超分辨率 最大后验概率 共轭梯度法 Gibbs模型 

分 类 号：TN911.73[电子电信—通信与信息系统]