基于聚类的非共线多CCD遥感图像误匹配点去除方法

发布时间：2018-07-09 11:13

  本文选题：遥感图像 + 非共线　； 参考：《电子与信息学报》2017年10期

【摘要】：针对非共线多CCD遥感图像匹配点的分布特点,该文提出一种基于聚类的误匹配点去除方法。首先,根据匹配点的沿轨方向偏移量曲线,获取匹配点的多维特征向量。然后,对匹配点集进行聚类处理,将所有点聚为一个类簇,最后根据簇半径序列曲线的变化趋势区分正确点和误匹配点。通过天绘1号02星全色遥感图像的实验和处理,结果表明在误匹配点去除和正确匹配点保留方面所提算法与其它方法相比具有更好性能。
[Abstract]:According to the distribution of matching points in non-collinear multi-CCD remote sensing images, this paper proposes a method for removing mismatched points based on clustering. Firstly, the multi-dimensional eigenvector of the matching point is obtained according to the offset curve of the matching point along the orbit direction. Then, the matching point set is clustered and all the points are clustered into one cluster. Finally, the correct points and the mismatched points are distinguished according to the changing trend of the cluster radius sequence curve. Through the experiment and processing of the panchromatic remote sensing image of Tianyi 1 02 star, the results show that the proposed algorithm has better performance than other methods in the aspects of mismatch point removal and correct matching point retention.
【作者单位】： 中国科学院空间信息处理与应用系统技术重点实验室;中国科学院电子学研究所;中国科学院大学;
【基金】：国家自然科学基金(41001214)~~
【分类号】：TP751
，

本文编号：2109171