几何校正重采样分配算法

Distribution algorithm in geometric correction resampling

在根据解析空中三角测量原理计算出地面实际坐标值进行遥感图像几何校正的过程中,往往需要重采样的操作。常规的利用一定大小搜索框历遍全图来完成重采样的方法对计算机硬件要求很高,需要大量的内存和时间来进行计算。作者提出了一种新的的算法:运用逆反思路,对原图像元进行分配来达到搜索框算法同样的效果。在相同大图像处理的情况下,该算法效率远远超过了常规算法。

In the process of remote sensing image＇s geometric correction based on analytic aerotriangulation, we always resample image by pixels＇ ground coordinates. The traditional resample method is searching though whole im- age by a certain size of kernel. This kernel method needs much computer memory and time to complete tasks. The author of this article brings forward a new method in reverse idea, which distribute each raw image pixel in order to get the same result of kernel method but save much computing memory and time. In the case of big image, this method is much more effective than traditional methods.