测绘通报 ›› 2021, Vol. 0 ›› Issue (9): 108-111.doi: 10.13474/j.cnki.11-2246.2021.0284

• 技术交流 • 上一篇    下一篇

高光谱遥感异常目标探测方法性能对比分析

王德强, 赵德良   

  1. 山东省地质测绘院, 山东 济南 250002
  • 收稿日期:2020-09-15 修回日期:2020-11-03 出版日期:2021-09-25 发布日期:2021-10-11
  • 通讯作者: 赵德良。E-mail:99755738@qq.com
  • 作者简介:王德强(1972-),男,高级工程师,主要从事测绘地理信息、摄影测量、遥感等方面的研究。E-mail:645938796@qq.com
  • 基金资助:
    山东省地质测绘院科研项目(KY201907)

Comparative analysis of hyperspectral remote sensing abnormal target detection methods

WANG Deqiang, ZHAO Deliang   

  1. Shandong Institute of Geological Surveying and Mapping, Jinan 250002, China
  • Received:2020-09-15 Revised:2020-11-03 Online:2021-09-25 Published:2021-10-11

摘要: 本文利用高光谱遥感异常目标探测理论,探讨了目前已有的几种异常目标探测方法;通过对3组数据进行试验,并从探测率和虚警率、ROC曲线及其下的面积及算法的运行时间,对几种异常检测算法的检测性能进行对比分析;最后基于统计模型和基于表示模型对3组数据的检测效果进行对比分析,从而得出适合于不同数据的检测方法,为高光谱遥感异常目标探测提供更多的支持。

关键词: 高光谱遥感, 异常检测, 对比分析, 探测率, 统计模型

Abstract: This paper is based on the theory of hyperspectral remote sensing abnormal target detection, mainly explores several existing abnormal target detection methods, through the experimental comparison and analysis of the detection performance of each algorithm, in order to get a better detection method, and provide more support for hyperspectral remote sensing abnormal target detection. In this paper, through the experiment of three groups of data, and from the detection rate and false alarm rate, the area under the ROC curve and the running time of the algorithm, the detection performance of the above-mentioned anomaly detection algorithms is compared and analyzed. Finally, the detection effect of the algorithm based on the statistical model and the representation model is compared and analyzed, so as to obtain the more suitable for different data Detection method.

Key words: hyperspectral remote sensing, anomaly detection, comparative analysis, detection rate, statistical model

中图分类号: