电子学报 ›› 2017, Vol. 45 ›› Issue (2): 337-345.DOI: 10.3969/j.issn.0372-2112.2017.02.011

• 学术论文 • 上一篇    下一篇

基于微分搜索的高光谱图像非线性解混算法

陈雷1,2,3, 郭艳菊4, 葛宝臻1,3   

  1. 1. 天津大学精密仪器与光电子工程学院, 天津 300072;
    2. 天津商业大学信息工程学院, 天津 300134;
    3. 光电信息技术教育部重点实验室, 天津 300072;
    4. 河北工业大学电子信息工程学院, 天津 300401
  • 收稿日期:2015-10-13 修回日期:2016-05-26 出版日期:2017-02-25 发布日期:2017-02-25
  • 作者简介:陈雷,男,1980年2月生于河北唐山.现为天津大学精密仪器与光电子工程学院博士后,天津商业大学信息工程学院副教授.主要研究方向为高光谱图像处理、仿生智能计算、盲信号处理等.E-mail:chenleitjcu@139.com;郭艳菊,女,1980年8月生于河北邢台.博士,讲师,研究方向为高光谱图像处理、仿生智能计算等;葛宝臻,男,1964年10月生于内蒙古卓资,教授,博士生导师,主要研究方向为光电成像技术,激光粒子测量.
  • 基金资助:

    国家自然科学基金(No.61401307);中国博士后科学基金(No.2014M561184);天津市应用基础与前沿技术研究计划项目(No.15JCYBJC17100)

Nonlinear Unmixing of Hyperspectral Images Based on Differential Search Algorithm

CHEN Lei1,2,3, GUO Yan-ju4, GE Bao-zhen1,3   

  1. 1. School of Precision Instrument and Opto-Electronics Engineering, Tianjin University, Tianjin 300072, China;
    2. School of Information Engineering, Tianjin University of Commerce, Tianjin 300134, China;
    3. Key Laboratory of Opto-Electronic Information Technology, Ministry of Education, Tianjin 300072, China;
    4. School of Electronic and Information Engineering, Hebei University of Technology, Tianjin 300401, China
  • Received:2015-10-13 Revised:2016-05-26 Online:2017-02-25 Published:2017-02-25

摘要:

针对线性混合模型在实际高光谱图像解混过程中的局限性,提出一种新的基于微分搜索的非线性高光谱图像解混算法.在广义双线性模型的基础上采用重构误差作为解混的目标函数,将非线性解混问题转化为最优化问题.将目标函数中的待求参数映射为微分搜索过程中的位置变量,利用微分搜索算法对目标函数进行优化求解.在求解过程中,通过执行搜索范围控制等机制满足高光谱图像解混的约束要求,进而求得丰度系数和非线性参数,实现非线性高光谱图像解混.仿真数据和真实遥感数据实验结果表明,所提出的非线性解混算法可以有效克服线性模型下解混算法的局限性,避免了由于使用梯度类优化方法而易陷入局部收敛的问题,较之其它高光谱图像解混算法具有更好的解混精度.

关键词: 高光谱图像, 谱解混, 非线性模型, 群智能优化, 微分搜索算法

Abstract:

A novel nonlinear hyperspectral image unmixing algorithm based on differential search is proposed for solving the limitations of linear mixing model.The reconstruction error is used as the objective function for unmixing based on generalized bilinear model and the nonlinear unmixing is transformed into the optimization problem.The parameters in objective function are mapped onto the location variables of the search process and the differential search algorithm is used to optimize the objective function.In the optimization process,the constraint conditions for hyperspectral image unmixing are fulfilled by implementing the search range controlling strategy.And then,the abundance and the nonlinear parameters for unmixing can be obtained.Experiments on synthetic data and real data validate that the proposed nonlinear unmixing algorithm can effectively overcome the limitations of linear unmixing algorithm,as well as the local convergence of gradient optimization method,and the performance of the proposed algorithm is better than other state-of-the-art hyperspectral image unmixing algorithms.

Key words: hyperspectral images, spectral unmixing, nonlinear model, swarm intelligence optimization, differential search algorithm

中图分类号: