一种优化的基于倾斜刃边的点扩散函数重建方法  

An Optimized Reconstruction Algorithm for Point Spread Function Based on Slant Step Edge

作　　者：曲梦雅 张永生[1] 李润生[1] QU Mengya ZHANG Yongsheng LI Runsheng(Zhengzhou Institute of Surveying and Mapping, Zhengzhou 450052, Chin)

机构地区：[1]郑州测绘学院,河南郑州450052

出　　处：《测绘科学技术学报》2016年第6期611-616,共6页Journal of Geomatics Science and Technology

基　　金：国家自然科学基金项目(41001262)

摘　　要：已有的倾斜刃边法受边缘亚像素定位精度、边缘扩散函数(ESF)样本质量及ESF曲线拟合方法的限制,点扩散函数(PSF)重建的稳定性和精度不高。为此,提出了一种优化的基于倾斜刃边的PSF估计方法,将梯度算子引入刃边直线拟合,对刃边边缘进行了高精度的亚像素位置修正;并采用基于移动窗口的ESF去噪及重采样方法精化ESF样本质量;最后通过高斯函数拟合得到稳健的PSF估计值。实验结果表明,改进算法的边缘直线拟合精度优化效果明显,且PSF重建精度较高,稳定性强。The existing slant step edge methods, greatly influenced by the edge＇s sub-pixel positioning precision, the Edge Spread Function （ESF） samples＇ quality and fitting models of the ESF curve, do not perform well in sta- bility and accuracy. An optimized reconstruction algorithm for Point Spread Function （PSF） with better precision is proposed using the slant step edge. The gradient operator is introduced in edge line fitting to refine the edge points＇ sub-pixel positions for greater accuracy. The improved method adopts new techniques of moving window in the ESF de-noising and resampling method to improve ESF sample＇s quality. Finally, robust PSF reconstruction results are obtained by Gaussian function fitting. Experiments demonstrate that the new algorithm shows great performance gain in the precision of step edge line fitting and also the superiority in accuracy and stability of the PSF reconstruction.

关 键 词：点扩散函数 边缘扩散函数 倾斜刃边法 梯度算子 亚像素拟合 

分 类 号：P237[天文地球—摄影测量与遥感]