测绘通报 ›› 2017, Vol. 0 ›› Issue (12): 82-85.doi: 10.13474/j.cnki.11-2246.2017.0384

• 学术研究 • 上一篇    下一篇

基于形状与纹理特征的高分辨率SAR影像桥梁目标自动提取方法

贾洪果, 刘路遥, 韦博文, 张梦意, 武燕, 章昊   

  1. 西南交通大学地球科学与环境工程学院测绘遥感信息系, 四川 成都 610031
  • 收稿日期:2017-08-22 出版日期:2017-12-25 发布日期:2018-01-05
  • 作者简介:贾洪果(1980-),女,博士,讲师,研究方向为合成孔径雷达干涉测量。E-mail:lemon_gg@hotmail.com
  • 基金资助:
    国家重点研发计划资助(2017YFB0502700);国家自然科学基金(41701535);卫星测绘技术与应用国家测绘地理信息局重点实验室经费(KLSMTA-201602)

Automatic Extraction of Bridges with Shape and Texture Characteristics Using High Resolution SAR Images

JIA Hongguo, LIU Luyao, WEI Bowen, ZHANG Mengyi, WU Yan, ZHANG Hao   

  1. Faculty of Geosciences and Environmental Engineering, Southwest Jiaotong University, Chengdu 610031, China
  • Received:2017-08-22 Online:2017-12-25 Published:2018-01-05

摘要: 基于高分辨率SAR影像中桥梁目标特征的分析,提出了一种桥梁目标自动提取方法。首先使用多影像时序平均法对SAR影像进行预处理,去除SAR影像噪声;再使用阈值法对水体分割;最后采用形状特征与纹理特征相结合的方法提取桥梁目标。本方法在提取效果上明显优于传统桥梁目标提取方法,其提取准确率可达87.5%。证明本文方法在面向高分辨率雷达影像的桥梁目标自动提取应用上的可靠性和可行性。

关键词: 高分辨率, SAR影像, 桥梁, 形状特征, 纹理特征, 自动提取

Abstract: This study proposed an automatic extraction method of bridge information based on the shape and texture characteristics of bridges in high resolution SAR images.This method consists of three steps:①elimination of noise by time-series image averaging;②detection of water-areas by a threshold;③extraction of bridge information based shape and texture characteristics of bridges.The experimental results show that the precise of extraction by the method proposed in this study can reach 87.5% and is significantly better than traditional methods.The experimental evaluation indicated that this method has high reliability and feasibility for automatic extraction of bridge information from high resolution SAR images.

Key words: high resolution, SAR images, bridge, shape characteristics, texture characteristics, automatic extraction

中图分类号: