测绘学报 ›› 2015, Vol. 44 ›› Issue (3): 274-281.doi: 10.11947/j.AGCS.2015.20140257

• 摄影测量学与遥感 • 上一篇    下一篇

多基线机载合成孔径雷达影像匹配的SANCC法

丁昊1,2, 张继贤2, 黄国满2, 朱建军1   

  1. 1. 中南大学地球科学与信息物理学院, 湖南 长沙 410083;
    2. 中国测绘科学研究院, 北京 100830
  • 收稿日期:2014-05-23 修回日期:2014-11-02 出版日期:2015-03-20 发布日期:2015-04-01
  • 作者简介:丁昊(1986—),女,博士生,研究方向为雷达摄影测量. E-mail:daisyaza@qq.com
  • 基金资助:
    测绘地理信息公益性行业科研专项(201412002;201412010);中国测绘科学研究院基本科研项目(7771303)

Multi-image Matching of Airborne SAR Imagery by SANCC

DING Hao1,2, ZHANG Jixian2, HUANG Guoman2, ZHU Jianjun1   

  1. 1. School of Geosciences and Info-Physics, Central South University, Changsha 410083, China;
    2. Chinese Academy of Surveying and Mapping, Beijing 100830, China
  • Received:2014-05-23 Revised:2014-11-02 Online:2015-03-20 Published:2015-04-01
  • Supported by:
    Public Science Research Program of Surveying,Mapping and Geoinformation(Nos.201412002;201412010);Basic Scientific Research Fund of Chinese Academy of Surveying and Mapping(No.7771303)

摘要: 为了有效利用合成孔径雷达(synthetic aperture radar,SAR)多基线影像的几何信息和辐射信息,提高匹配精度,提出了一种适用于多基线SAR幅度影像的自适应归一化互相关系数和(sum of adaptive normalized cross-correlation,SANCC)影像匹配方法.该方法首先利用SAR成像参数、平台参数和物方高程范围构建匹配方向线;然后引入Gestalt原理的接近性和相似性原则对窗口内的像素加权,计算获得沿匹配方向线的SANCC值;最后采用winner-take-all(WTA)优化策略获取多基线影像的匹配结果和物方三维信息.利用国产机载SAR系统获取的多基线影像进行匹配试验,结果表明与常规的相关匹配方法相比,该方法可以获得更为密集、精确的匹配点,有效减少了由重复纹理造成的误匹配,并能较好地解决由纹理匮乏导致的匹配难题.

关键词: 自适应归一化互相关系数和, 多基线影像匹配, 匹配方向线, 机载SAR影像

Abstract: In order to improve accuracy of SAR matching, a multi-image matching method based on sum of adaptive normalized cross-correlation (SANCC) is proposed. It utilizes geometrical and radiometric information of multi-baselinesynthetic aperture radar (SAR)images effectively. Firstly, imaging parameters, platform parameters and approximate digital surface model (DSM) are used to predict matching line. Secondly, similarity and proximity in Gestalt theory are introduced to SANCC, and SANCC measures of potential matching points along the matching line are calculated. Thirdly, multi-image matching results and object coordinates of matching points are obtained by winner-take-all (WTA) optimization strategy. The approach has been demonstrated with airborne SAR images acquired by a Chinese airborne SAR system (CASMSAR system). The experimental results indicate that the proposed algorithm is effective for providing dense and accuracy matching points, reducing the number of mismatches caused by repeated textures, and offering a better solution to match in poor textured areas.

Key words: SANCC, multi-image matching, matching line, airborne SAR imagery

中图分类号: