可利用任意形状刃边的遥感器在轨点扩散函数估算  

Estimation of the on-orbit point spread function for remote sensors applicable to any edge

作　　者：王俊琦[1,2] 张立国[1] 陈起行 付天骄[1] 

机构地区：[1]中国科学院长春光学精密机械与物理研究所,吉林长春130033 [2]中国科学院大学,北京100049

出　　处：《光电子．激光》2016年第5期549-556,共8页Journal of Optoelectronics·Laser

基　　金：国家“863”高技术研究发展计划(863-2-5-1-13B)资助项目

摘　　要：为了使空间相机在无靶标场景中也能提取系统点扩散函数(PSF,point spread function),基于倾斜刃边法,提出了一种利用任意形状刃边估计PSF的方法。通过分析刃边线扩散函数(LSF,line spread function)的特点,以任意形状刃边LSF的直接提取为基础对PSF进行估算。建立理想刃边模型进行实验,该方法能够有效地提取成像系统的PSF,并具有较高的测量精度与较好的抗噪性能。将本文方法应用在实际遥感场景中,分别利用倾斜靶标刃边和随机地物刃边估算PSF,以倾斜刃边法估算出的PSF为标准进行对比,误差分别为4.19%与5.09%;该方法应用于不同遥感场景时,估算出各自的PSF并用维纳滤波复原,复原后的图像质量有很大提升。该方法在保证PSF估算精度的同时有效减少了对靶标的依赖,具有良好的应用前景。In order to extract the point spread function （PSF） from more remote sensing scenes and avoid the problem that normal method usually depends on ground target, this paper proposes a PSF assess- ment algorithm based on the slanted-edge method, and applies it to the images with arbitrary edges. The feature of line spread function （LSF） is analyzed,and then the PSF estimation is realized by extracting the LSF from arbitrary edges. Ideal models of slanted-edge and curved,edge are established. The experi- ments show that the proposed method can estimate the PSF effectively, the errors are all below 1%, and it also has high precision and good performance in noise resistance. The method is applied in real remote sensing scenes,and the slanted-edge which is clipped from the ground target and the random edge in the same scene are used to assess the PSF. Compared with the PSF estimated by the slanted-edge method, the errors can reach 4. 19% and 5.09%. The PSF estimated by the proposed algorithm is applied to image restoration using the wiener filter,and the image qualities are enhanced obviously. The method in this paper ensures the precision of PSF estimation and reduces the reliance on ground target.

关 键 词：遥感 点扩散函数(PSF) 刃边法 弯曲刃边 图像复原 

分 类 号：TP79[自动化与计算机技术—检测技术与自动化装置]