测绘学报 ›› 2019, Vol. 48 ›› Issue (5): 609-617.doi: 10.11947/j.AGCS.2019.20170746

• 摄影测量学与遥感 • 上一篇    下一篇

极化SAR参数优化与光学波谱相结合的面向对象土地覆盖分类

赵诣, 蒋弥   

  1. 河海大学地球科学与工程学院, 江苏 南京 211100
  • 收稿日期:2017-12-27 修回日期:2018-07-29 出版日期:2019-05-20 发布日期:2019-06-05
  • 通讯作者: 蒋弥 E-mail:mijiang@hhu.edu.cn
  • 作者简介:赵诣(1995-),男,硕士生,研究方向为极化SAR数据处理与应用。E-mail:zyhhu@hhu.edu.cn
  • 基金资助:
    国家重点研发计划(2018YFC0407900);国家自然科学基金(41774003);江苏省自然科学基金(BK20171432);中央高校基本科研业务费专项资金资助(2018B17714)

Integration of SAR polarimetric parameters and multi-spectral data for object-based land cover classification

ZHAO Yi, JIANG Mi   

  1. School of Earth Sciences and Engineering, Hohai University, Nanjing 211100, China
  • Received:2017-12-27 Revised:2018-07-29 Online:2019-05-20 Published:2019-06-05
  • Supported by:
    The National Key Research and Development Program of China (No. 2018YFC0407900);The National Natural Science Foundation of China (No. 41774003);The Natural Science Foundation of Jiangsu Province (No. BK20171432);The Fundamental Research Funds for the Central Universities (No. 2018B17714)

摘要: 提出一种基于极化参数优化的面向对象分类方法。该方法结合光学和SAR数据,有效提高了对地物的识别能力。本文方法的关键在于:在H-A-α分解中,使用光学影像指导SAR影像选择同质点,使其更精确地估计极化参数并结合光学波谱信息作为输入特征;使用面向对象的分类方法,仅将光学影像作为分割输入,避免SAR噪声引起的分割错误。以美国Bakersfield地区的Sentinel-1/2数据为例,确定7种地物类型,对比分析不同输入与不同分类器对分类结果的影响。研究表明,优化输入参数在纹理丰富区域能够有效提高分类精度;面向对象的分类结果更加稳定并较好地维持地表几何特征;改进分类方法较传统分类方法总体精度提高了近10%,达到92.6%。

关键词: 合成孔径雷达, 极化, 多光谱, 数据融合, 面向对象, 土地覆盖分类

Abstract: An object-based approach is proposed for land cover classification using optimal polarimetric parameters. The ability to identify targets is effectively enhanced by the integration of SAR and optical images. The innovation of presented method can be summarized in the following two main points:① estimating polarimetric parameters (H-A-α decomposition) through optical image as a driver; ② a multi-resolution segmentation based on optical image only is deployed to refine classification results. The proposed method is verified by using Sentinel-1/2 datasets over Bakersfield area, California.The results are compared against those from pixel-based SVM classification using the ground truth from the National Land Cover Database (NLCD). A detailed accuracy assessment complied for seven classes of surfaces shows that the proposed method outperforms the conventional approach by around 10%, with an overall accuracy of 92.6% over regions with rich texture.

Key words: synthetic aperture radar(SAR), polarimetric, multispectral, data fusion, object-based, land-cover classification

中图分类号: