测绘学报 ›› 2022, Vol. 51 ›› Issue (7): 1091-1107.doi: 10.11947/j.AGCS.2022.20220070

• 院士论坛 • 上一篇    下一篇

多时相遥感影像的变化检测研究现状与展望

张祖勋1, 姜慧伟2, 庞世燕3, 胡翔云1,4   

  1. 1. 武汉大学遥感信息工程学院, 湖北 武汉 430079;
    2. 国家基础地理信息中心, 北京 100830;
    3. 华中师范大学人工智能教育学部, 湖北 武汉 430079;
    4. 湖北珞珈实验室, 湖北 武汉 430079
  • 收稿日期:2022-01-31 修回日期:2022-06-03 发布日期:2022-08-13
  • 通讯作者: 姜慧伟 E-mail:huiwei_jiang@whu.edu.cn
  • 作者简介:张祖勋(1937-),男,教授,博士生导师,中国工程院院士,研究方向为数字摄影测量与遥感。E-mail:zhangzx@cae.cn
  • 基金资助:
    湖北珞珈实验室专项基金(220100028)

Review and prospect in change detection of multi-temporal remote sensing images

ZHANG Zuxun1, JIANG Huiwei2, PANG Shiyan3, HU Xiangyun1,4   

  1. 1. School of Remote Sensing and Information Engineering, Wuhan University, Wuhan 430079, China;
    2. National Geomatics Center of China, Beijing 100830, China;
    3. Faculty of Artificial intelligence in Education, Central China Normal University, Wuhan 430079, China;
    4. Hubei Luojia Laboratory, Wuhan 430079, China
  • Received:2022-01-31 Revised:2022-06-03 Published:2022-08-13
  • Supported by:
    Special Fund of Hubei Luojia Laboratory (No. 220100028)

摘要: 变化检测作为摄影测量与遥感领域的研究热点之一,也是人工智能体系中极具研究价值的技术分支。二者的快速发展与深度融合,已使海量、复杂和多样的遥感数据快速智能化处理成为可能,广泛应用于资源监测、城市规划、灾害评估等诸多领域。随着遥感技术和计算能力的不断革新,变化检测体系也在不断发展和演化。本文主要从几何和语义两个角度对变化检测方法进行了分析和归纳总结,重点分析了几何信息的利用方式及深度神经网络的特征融合方式,随后总结了常用的变化检测公开样本数据集,最后对当前变化检测应用中遇到的核心问题及未来发展趋势进行了梳理与展望。

关键词: 摄影测量, 遥感影像, 多时相, 变化检测, 综述

Abstract: Change detection (CD), as one of the hot spots in the field of photogrammetry and remote sensing, is a technology branch with important research value in artificial intelligence (AI) systems. The rapid development and deep integration of the two subjects have made it possible to automatic and intelligent process the massive, complex and diverse remote sensing datasets. It has been widely used in many areas, such as resources monitoring, urban planning, disaster assessment, etc. With the development of the remote sensing and computing power, change detection is also evolving. This paper mainly analyzes and summarizes the current change detection methods from the perspectives of geometry and semantics. We focus on the utilization of geometric information and the feature fusion method of deep neural networks. Furthermore, the paper provides a summary of open datasets derived from different sensors, along with information related to change detection. Finally, we point out the challenges of the change detection and the future trends of the area are summarized and prospected.

Key words: photogrammetry, remote sensing images, multi-temporal, change detection, review

中图分类号: