地震地质

• •    

基于特征点的SAR影像配准方法研究

刘智勇1,来立永2,张耿斌1,祁宏昌1,潘屹峰3,彭林才2,吴希文2   

  1. 1. 广东电网有限责任公司广州供电局
    2. 广东工业大学
    3. 中科云图科技有限公司
  • 收稿日期:2021-04-07 修回日期:2021-07-05 发布日期:2021-09-26
  • 通讯作者: 彭林才
  • 基金资助:
    广州市科技计划项目;中国南方电网有限责任公司广州供电局有限公司重点科技项目;广州市珠江新星项目;广东省科学院实施创新驱动发展能力建设专项

Research on SAR Image Coregistration Method Based on Feature Points

  • Received:2021-04-07 Revised:2021-07-05 Published:2021-09-26
  • Contact: Lincai Peng

摘要: 合成孔径雷达(Synthetic Aperture Radar, SAR)影像配准是合成孔径雷达干涉测量 (Interferometric SAR, InSAR)数据处理的一个关键步骤,是生成良好干涉图的必要条件,但是由于各种因素,导致精确的配准非常困难。针对常规的配准方法计算量大、误匹配率高以及配准精度差的问题,我们提出了一种利用特征点进行配准的方法。首先,使用加速鲁棒特征算子Speeded Up Robust Feature (SURF)从主强度图中提取特征点,然后结合外部地理数据剔除不当的特征点,并对剩下的特征点进行互相关处理,最后通过多项式拟合计算主辅影像之间的转换模型。通过珠三角洲地区与智利地震的SAR影像的配准实验结果表明,与标准的配准方法相比,本文提出的方法具有配准效率高、可靠性强的优点。

关键词: 配准, SURF, 特征点

Abstract: Synthetic aperture radar (SAR) image coregistration is a key step in synthetic aperture radar interferometric(InSAR) data processing. It is a necessary condition for generating good interferograms. However, due to various factors, accurate coregistration is very difficult. Aiming at the problems of large amount of calculation, high mismatch rate and poor coregistration accuracy of conventional coregistration methods, we propose a coregistration method using feature points. First, Speeded Up Robust Feature (SURF) operator is used to extract the feature points from the master amplitude map, and then combined with external geographic data to remove improper feature points and turn them into the remaining feature points for cross-correlation processing, and finally calculating the coregistration of the main and slave images through multiple formulas through the SAR image in the Pearl River Delta and Chile earthquake. The experimental results show that, compared with the standard coregistration method, the method proposed here has the advantages of high coregistration efficiency and strong reliability.

Key words: coregistration, Speeded Up Robust Feature, Feature Points