遥感图像配准中特征点选择的高性能算法研究及其实现

High Performance Algorithm of Selecting Characteristic Points in the Remote Sensing Images Registration and Its Implementation

在基于特征点的遥感图像配准过程中 ,特征点的自动、快速提取是影响配准精度和速度的关键 .采用手工选择特征点 ,能够保证特征点的分布 ,但是自动化程度低而且耗时 .该文提出了一种用于解决特征点选择的高性能算法 .该算法选择的特征点分布均匀 ,保证了配准的精度 ;特征点的选择过程完全自动化 ,并通过并行优化进一步提高了算法的效率 .

Selecting the characteristic points automatically and fast is one of the critical steps when the image registration for remote sensing is implemented based on characteristic points. Characteristic points selected by hand distribute evenly, but it is time-consuming and lacks the automatization. This paper puts forward a high performance algorithm to select characteristic points based on the entropy. It improves the accuracy of registration. The efficiency is improved by the parallel technique.