基于多项式变换的遥感图像几何校正并行算法的研究与实现

研究和实现高效的几何枝正并行算法，能够有效地提高遥感图像预处理的工作效率。本文首先研究了基于多项式变换的几何校正算法，然后提出了新的并行几何校正算法。该算法通过局部边界计算达到延迟通信的目的，很好地解决了计算局部性问题；在机群系统上实现后，对多幅大小不同的G比特遥感数据集进行了测试，获得了正确的结果和良好的并行性能。

Investigation and Realization of Parallel Algorithms for the Geometric Correction of Remote Sensing Images Based on Polynomial Transform

To study and realize effective parallel algorithms for geometrical correction can improve the job efficiency of remote sensing image preprocessing. Firstly, the geometrical correction based on polynomial transform is studied. Then a new parallel algorithm is proposed, which solves the problem of computing locality through calculating the local data edges to delay required communications. Finally, this parallel algorithm is realized on a cluster system, and several remote sensing images with the size of gigabytes are used to test the algorithm performance. The results show the validity and good efficiency of our algorithm.