应用科学学报 ›› 2010, Vol. 28 ›› Issue (4): 354-360.doi: 10.3969/j.issn.0255-8297.2010.04.005

• 信号与信息处理 • 上一篇    下一篇

从高分辨率遥感图像中自动识别水上桥梁

彭碧发1;2, 谭毅华1;2, 陶超1;2, 田金文1;2   

  1. 1. 华中科技大学图像识别与人工智能研究所,武汉430074
    2. 华中科技大学多谱信息处理技术国家级重点实验室,武汉430074
  • 收稿日期:2010-04-22 修回日期:2010-05-21 出版日期:2010-07-23 发布日期:2010-07-23
  • 通信作者: 谭毅华,副教授,研究方向:图像/视频压缩和分析、智能视觉监控、遥感图像目标检测和识别,E-mail: yihua.tan@263.net
  • 基金资助:

    国家“863”高技术研究发展计划基金(No.2007AA12Z153)资助

     

Automatic Detection of Bridge over Water in High-Resolution Remote-Sensing Image


PENG Bi-fa1;2, TAN Yi-hua1;2, TAO Chao1;2, TIAN Jin-wen1;2   



  1. 1. Institute for Pattern Recognition and Artificial Intelligence, Huazhong University of Science and Technology,
    Wuhan 430074, China
    2. National Key Laboratory of Science and Technology on Multi-spectral Information Processing,
    Huazhong University of Science and Technology, Wuhan 430074, China
  • Received:2010-04-22 Revised:2010-05-21 Online:2010-07-23 Published:2010-07-23

摘要:

提出一种高分辨率遥感图像中水上桥梁的自动识别方法. 根据水域的均质性将图像分割成水域和非水域两部分,扫描非水域部分对桥梁进行粗提取. 以粗提取所获得的区域中心为起始点进行区域增长,用全方向扫描获取临界区域,进而计算桥梁方向. 最后沿桥梁方向往桥梁两边延伸,获得精确的桥梁区域. 实验表明该方法可以自动识别出高分辨率遥感图像中的水上桥梁目标,精确度高,鲁棒性好.

关键词: 高分辨率, 水上桥梁, 区域增长, 自动识别

Abstract:

An automatic detection method to identify bridges over water in high-resolution remote-sensing images is proposed. The input image is first segmented into two water and non-water areas according to homogeneity of the water area. The image is scanned, and the bridge is extracted from the non-water area. The orientation of the bridge is determined using region growing and multiple-direction scanning. The bridge is then accurately outlined from the coarse extraction by extension along its orientation. Experimental results show that the proposed method can automatically detect bridge over water in high-resolution remote-sensing images. Performance evaluation indicates that the proposed method is robust and accurate.

Key words: high-resolution, bridge over water, region growing, automatic recognition

中图分类号: