加权SUSAN算法与遥感图像矢量化

Weighted SUSAN Algorithm and Vectorization of Remote Sensing Image

本文对经典SUSAN算法进行了改进，提出了WSUSAN，即加权SUSAN的概念，与经典算法相比，WSUSAN角点的检出精度大幅提高，而运算复杂度和运算量仍保持在原数量级。作为WSUSAN的应用，本文进行了基于WSUSAN角点的遥感图像矢量化的研究，主要研究的内容包括过检出角点的筛选、角点排序和矢量化等。经过仿真测试，证明该算法有很好的鲁棒性。

A new Algorithm, Weighted SUSAN （WSUSAN）,which is an improvement of classical SUSAN algorithm focusing on corner detection,is put forward Comparing with the classical algorithm,the new one has a significant improvement in accuracy, while keeps the efficiency in the same level.As an application of WSUSAN,A research of remote sensing image vectorization based on WSUSAN corner detection result is brought about. The research mainly includes：the filtering of over-detected corners,the sort of corners and vectorization.It is proved that the algorithm has a good robustness base on the simulation.