测绘通报 ›› 2019, Vol. 0 ›› Issue (3): 61-66.doi: 10.13474/j.cnki.11-2246.2019.0079

• 学术研究 • 上一篇    下一篇

基于随机森林的FY-2G云检测方法

付华联1,2, 冯杰1, 李军1, 刘军2   

  1. 1. 成都理工大学, 四川 成都 610000;
    2. 中国科学院深圳先进技术研究院, 广东 深圳 518055
  • 收稿日期:2018-04-18 出版日期:2019-03-25 发布日期:2019-04-02
  • 通讯作者: 李军。E-mail:kingleo_cn@qq.com E-mail:kingleo_cn@qq.com
  • 作者简介:付华联(1993-),女,硕士生,研究方向为图像处理与模式识别。E-mail:1481671930@qq.com
  • 基金资助:
    国家重点研发计划(2017YFB0504203);深圳国际合作研究项目(GJHZ20160229194322570);广东省科技计划(2017A050501027);中科院西部之光项目(2016-QNXZ-A-5);深圳基础研究项目(JCYJ20160429191127529)

Cloud detection method of FY-2G satellite images based on random forest

FU Hualian1,2, FENG Jie1, LI Jun1, LIU Jun2   

  1. 1. Chengdu University of Technology, Chengdu 610000, China;
    2. Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, Shenzhen 518055, China
  • Received:2018-04-18 Online:2019-03-25 Published:2019-04-02

摘要: 根据遥感影像中云检测原理,提出了基于随机森林的遥感影像云检测方法,并将其应用于FY-2G影像。结合国家气象卫星中心(NSMC)的云检测产品数据进行了算法的精度检验,云检测个例的精度检验结果显,最高命中率(POD)为88.32%,最低误报率(FAR)为9.36%,临界成功指数(CSI)为80.14%。结果表明,该方法有效地提高了云检测精度,同时能正确标识NSMC中部分误判的情况。

关键词: 云检测, 国家气象卫星中心, 随机森林, 大津法, FY-2G

Abstract: According to the principle of cloud detection in remote sensing images,a method based on the random forest was proposed in this paper and applied to FY-2G images.Combining the cloud inspection product of the National Meteorological Satellite Center (NSMC) for accuracy testing of the algorithm,the accuracy test of the cloud detection shows that the Probability of Detection (POD) is 88.32%,the minimum False Alarm Rate (FAR) is 9.36%,and the Critical Success Index (CSI) is 80.14%.The results show that the method can improve the accuracy of cloud detection effectively,and identify the partial misjudgment in NSMC correctly.

Key words: cloud detection, NSMC, random forest, Otsu, FY-2G

中图分类号: