测绘学报 ›› 2019, Vol. 48 ›› Issue (12): 1542-1550.doi: 10.11947/j.AGCS.2019.20190453

• 综述 • 上一篇    下一篇

航摄影像密集匹配的研究进展与展望

袁修孝1, 袁巍1,2, 许殊1,3, 纪艳华1   

  1. 1. 武汉大学遥感信息工程学院, 湖北 武汉 430079;
    2. 东京大学空间信息科学中心, 东京 柏市 277-6568;
    3. 中国科学院空天信息创新研究院, 北京 100094
  • 收稿日期:2019-10-31 修回日期:2019-11-20 发布日期:2019-12-24
  • 作者简介:袁修孝(1963-),男,博士,教授,博士生导师,主要研究航空航天遥感高精度对地目标定位理论与方法、高分辨率卫星遥感影像几何处理等。E-mail:yuanxx@whu.edu.cn
  • 基金资助:
    国家自然科学基金(41771479);国家高分专项(民用部分)(50-H31D01-0508-13/15)

Research developments and prospects on dense image matching in photogrammetry

YUAN Xiuxiao1, YUAN Wei1,2, XU Shu1,3, JI Yanhua1   

  1. 1. School of Remote Sensing and Information Engineering, Wuhan University, Wuhan 430079, China;
    2. Center for Spatial Information Science, University of Tokyo, Kashiwa 2776568, Japan;
    3. Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100094, China
  • Received:2019-10-31 Revised:2019-11-20 Published:2019-12-24
  • Supported by:
    The National Natural Science Foundation of China (No. 41771479);The National High-Resolution Earth Observation System (the Civil Part) (No. 50-H31D01-0508-13/15)

摘要: 给出了航摄影像密集匹配的总体流程,依据是否显式使用光滑假设将密集匹配方法分为局部最优密集匹配和全局最优密集匹配两类,深入探讨了两种方法的关键技术,指出了从理论、技术、普适性和实用性方面值得关注的问题,期望能对相关研究有所裨益。

关键词: 航摄影像, 密集影像匹配, 局部最优匹配, 全局最优匹配, 光流场法, 深度学习方法

Abstract: The general workflow for dense matching of aerial images is given in this paper. Dense matching is divided into two categories, namely, those that utilize local matching algorithms and global matching algorithms, respectively. The key technologies of the two methods are analyzed in details. Concerns in theory, technology, universality and practicability are proposed. We hope it will be helpful to the related research on dense matching.

Key words: aerial image, dense image matching, local matching algorithm, global matching algorithm, optical flow field-based method, deep learning-based method

中图分类号: