English

应用科学学报 ›› 2012, Vol. 30 ›› Issue (1): 82-88.doi: 10.3969/j.issn.0255-8297.2012.01.013

• 信号与信息处理 • 上一篇    下一篇

结合光谱解混的高光谱图像异常目标检测SVDD算法

成宝芝， 赵春晖， 王玉磊   

  1. 哈尔滨工程大学信息与通信工程学院，哈尔滨150001
  • 收稿日期:2011-05-26 修回日期:2011-09-10 出版日期:2012-02-09 发布日期:2012-01-30
  • 通信作者: 作者简介：成宝芝，博士生，研究方向：高光谱图像处理，E-mail: chengbaozhib09@hrbeu.edu.cn；赵春晖，教授，博导，研究方向：智能信息与图像处理、信号处理，E-mail: zhaochunhui@hrbeu.edu.cn
  • 作者简介:作者简介：成宝芝，博士生，研究方向：高光谱图像处理，E-mail: chengbaozhib09@hrbeu.edu.cn；赵春晖，教授，博导，研究方向：智能信息与图像处理、信号处理，E-mail: zhaochunhui@hrbeu.edu.cn
  • 基金资助:

    国家自然科学基金(No.61077079)；高等学校博士学科点专项科研基金(No.20102304110013)；哈尔滨市优秀学科带头人基
    金(No.2009RFXXG034)资助

SVDD Algorithm with Spectral Unmixing for Anomaly Detection in Hyperspectral Images

CHENG Bao-zhi, ZHAO Chun-hui, WANG Yu-lei   

  1. College of Information and Communication, Harbin Engineering University, Harbin 150001, China
  • Received:2011-05-26 Revised:2011-09-10 Online:2012-02-09 Published:2012-01-30

PDF

1503

被引次数

35

摘要：

异常目标检测是高光谱数据处理的重要应用之一. 传统方法采用支持向量数据描述(support vector datadescription, SVDD)检测异常目标而不考虑图像自身存在的背景干扰，检测概率较低. 该文提出一种新方法，将光谱解混技术引入到基于SVDD的异常检测问题中，实现高光谱图像复杂背景信息和目标信息的分离，使解混后的误差数据含有丰富的目标信息，抑制了背景干扰. 利用非线性SVDD将解混误差数据映射到高维特征空间，充分利用高光谱图像波段间的非线性统计特性，完成异常目标的检测. 仿真实验结果表明，该算法提高了异常目标的检测能力，降低了虚警率.

关键词: 光谱解混, 支持向量数据描述, 异常检测

Abstract:

 Without considering background interferences, the classical algorithm of support vector data description (SVDD) has relatively low detection probability in hyperspectral anomaly detection. To solve the problem, this paper presented a new algorithm based SVDD, which includes hyperspectral unmixing to separate
target information from complicated background clutter. After spectral unmixing, the error datum includes abundant target information while effectively suppresses the background interference. The error datum is then mapped into a high-dimensional feature space with nonlinear SVDD. By exploiting nonlinear information between the spectral bands of hyperspectral imagery, anomaly targets are detected. The results show that the proposed algorithm can improve detection performance and decrease false alarm probability.

Key words:  spectral unmixing, support vector data description (SVDD), anomaly detection

中图分类号: