电子学报 ›› 2022, Vol. 50 ›› Issue (2): 339-345.DOI: 10.12263/DZXB.20201283

• 学术论文 • 上一篇    下一篇

基于拉普拉斯算子的综合孔径辐射计成像算法

杨晓城1, 杨真乙1, 阎敬业2, 武林2, 蒋明峰1   

  1. 1.浙江理工大学信息学院,浙江 杭州 310018
    2.中国科学院国家空间科学中心,北京 100190
  • 收稿日期:2020-11-16 修回日期:2021-07-30 出版日期:2022-02-25 发布日期:2022-02-25
  • 作者简介:杨晓城 男,1988年12月出生于江西九江.浙江理工大学信息学院讲师.研究方向为综合孔径辐射计、图像处理.E-mail: yangxiaoch209@163.com
    杨真乙 男,1996年12月出生于浙江台州.浙江理工大学信息学院硕士研究生.研究方向为综合孔径辐射计、图像处理.
  • 基金资助:
    国家自然科学基金(61672466);浙江省自然科学基金(LY18D060009);浙江省自然科学基金-数理医学学会联合基金重点项目(LSZ19F010001);浙江省科技厅重点研发项目(2020C03060)

Imaging Algorithm of Synthetic Aperture Interferometric Radiometers Based on Laplace Operator

YANG Xiao-cheng1, YANG Zhen-yi1, YAN Jing-ye2, WU Lin2, JIANG Ming-feng1   

  1. 1.School of Information Science and Technology,Zhejiang Sci-Tech University,Hangzhou,Zhejiang 310018,China
    2.National Space Science Center,Chinese Academy of Sciences,Beijing 100190,China
  • Received:2020-11-16 Revised:2021-07-30 Online:2022-02-25 Published:2022-02-25

摘要:

成像反演是综合孔径辐射计的一项关键内容.然而,综合孔径辐射计成像反演是病态的反问题.虽然传统的正则化方法能有效克服其病态性,但依然存在较大的重构误差.为了进一步降低重构误差,提出一种基于拉普拉斯算子的综合孔径辐射计成像算法.与传统的单参数正则化不同,该算法通过引入拉普拉斯算子项,对差分亮温构造拉普拉斯混合正则化,并利用多维扩展的广义交叉验证准则选取2个正则化参数.基于L波段的FPIR样机进行了仿真分析,仿真结果表明:与改进的最小范数正则化和带限正则化相比,该算法均方根误差平均降低了30%,峰值信噪比平均提高了3 dB,证明了其有效性.

关键词: 被动微波遥感, 综合孔径辐射计, 正则化, 重构误差, 拉普拉斯算子

Abstract:

The imaging inversion is a key content of synthetic aperture interferometric radiometers(SAIRs). However, the imaging inversion in SAIRs is an ill-posed inverse problem. Although the traditional regularization methods can effectively overcome the ill-conditioned property, there is still a large reconstruction error. In order to further reduce the reconstruction error, an imaging algorithm of SAIRs based on Laplace operator is proposed. Different from traditional single-parameter regularization, the proposed algorithm constructs the Laplace hybrid regularization for differential brightness temperature by introducing a new term of Laplace operator. Moreover, two regularization parameters are selected by use of multi-dimensional extended generalized cross-validation criterion. The simulation is based on the prototype of full polarization interferometric radiometer(FPIR). The simulation results show that compared with improved minimum norm regularization and band-limited regularization, the root mean square error for the proposed algorithm has reduced by 30%, and the peak signal to noise ratio has increased by 3 dB, proving its effectiveness.

Key words: passive microwave remote sensing, synthetic aperture interferometric radiometer, regularization, reconstruction error, Laplace operator

中图分类号: