Landsat-7缝隙数据恢复的算法研究  

Research on Algorithms for Recovering Landsat-7 Gap Data

作　　者：田晓红[1] 林友明[1] 

机构地区：[1]中国科学院中国遥感卫星地面站

出　　处：《计算机仿真》2007年第12期59-61,共3页Computer Simulation

摘　　要：扫描行校正器异常造成了Landsat-7图像数据丢失(称之为缝隙数据),由于数据本身仍保持了良好的辐射和几何特性,具有较好的可用性,因此研究缝隙数据的恢复算法将具有较好的理论价值和应用前景。目的就是通过仿真实验寻找一种较好的恢复算法。首先介绍了课题背景以及现有的算法——全局直方图匹配法和局部直方图匹配法,并在局部直方图匹配算法的基础上提出了自适应局部回归匹配算法。最后对同一幅缝隙图像用这三种算法来实现恢复,通过比较均方根误差和运行速度,论证了自适应局部回归匹配算法在精度方面要优于全局直方图匹配算法和局部直方图匹配算法,算法复杂度要优于局部直方图匹配算法,具有很高的可行性。For the data loss induced by the SLC＇s anomaly（called gap data）,as the data itself preserves good radicalization and geometry performances and the usability,it has good theoretical value and application foreground to research the algorithms for recovering the gap data.The purpose of this paper is to find out a better recovering algorithm.This paper first introduces the background of the problem and the existing algorithms such as global histogram match,local histogram match and then proposes a new algorithm—— adaptive local regression match.At last three algorithms are used to recover the same gap image and the best one is chosen by comparing the RMSE and complexity.It is testified that the adaptive local regression match is better than global histogram match and local histogram match on RMSE,especially better than local histogram match on complexity and has high usibility.

关 键 词：直方图 匹配 缝隙 填充 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]