Analysis of Posterior Probability Uncertainty for Classification of Hyperspectral Images by SupportVector Machines

被引:0
作者
Sun, Xiaoxia [1 ]
Li, Liwei [1 ]
Zhang, Bing [1 ]
Yang, Ling
机构
[1] Chinese Acad Sci, Lab Digital Earth, Beijing, Peoples R China
来源
PROCEEDINGS OF THE 2013 THE INTERNATIONAL CONFERENCE ON REMOTE SENSING, ENVIRONMENT AND TRANSPORTATION ENGINEERING (RSETE 2013) | 2013年 / 31卷
关键词
SVM; hyperspectral; posterior probability; classification; uncertainty; NEURAL-NETWORK;
D O I
暂无
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
This paper analyzes the uncertainty of classification posterior probability of support vector machine (SVM) using urban hyperspectral images. The hyperspectral images in Zhangye are selected as the study zone, and the sample parameter data were acquired based on the high resolution images and the ground survey information, the images were classified with parameter-optimized SVM to obtain the posterior probability graph for each class, and the posterior probability graphs were truncated using the threshold values of 0.2, 0.4, 0.6, 0.8 and 0.9 for analysis of the accuracy change of ground object classification at different probabilities. The results show that with the increase of truncation probability, the user accuracy in the classification results increases continuously, while the producer accuracy shows a declining tendency, and the overall classification accuracy also shows a declining tendency. The analysis of the posterior probability distribution of various types of ground objects shows that it is difficult to distinguish the posterior probability of some mixed ground objects. The untrained water body targets can be easily distinguished by the truncation probability, but the posterior probabilities of untrained red materials and white materials are mixed together. This shows that there exist some conditions in which the posterior probability of optimized SVM can not directly and effectively indicate the distinction of ground objects. The posterior probability should be used optionally, and at the same time, it is necessary to construct a more robust calculation method for the posterior probability.
引用
收藏
页码:22 / 25
页数:4
相关论文
共 50 条
  • [41] Adaptive Classification for Hyperspectral Image Data Using Manifold Regularization Kernel Machines
    Kim, Wonkook
    Crawford, Melba M.
    IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, 2010, 48 (11): : 4110 - 4121
  • [42] Survey of supervised classification techniques for hyperspectral images
    Qiu, Qingchen
    Wu, Xuelian
    Liu, Zhi
    Tang, Bo
    Zhao, Yuefeng
    Wu, Xinyi
    Zhu, Hongliang
    Xin, Yang
    SENSOR REVIEW, 2017, 37 (03) : 371 - 382
  • [43] Thematic classification with support subspaces in hyperspectral images
    Fursov, Vladimir Alekseyevich
    Bibikov, Sergey Alekseyevich
    Zherdev, Denis Alekseevich
    Kazanskiy, Nikolay Lvovich
    INTERNATIONAL JOURNAL OF ENGINEERING SYSTEMS MODELLING AND SIMULATION, 2020, 11 (04) : 186 - 193
  • [44] A NOVEL SCHEME FOR THE COMPRESSION AND CLASSIFICATION OF HYPERSPECTRAL IMAGES
    Xie, Bei
    Bose, Tamal
    Merenyi, Erzsebet
    2009 FIRST WORKSHOP ON HYPERSPECTRAL IMAGE AND SIGNAL PROCESSING: EVOLUTION IN REMOTE SENSING, 2009, : 287 - +
  • [45] Cooperative evolutionary classification algorithm for hyperspectral images
    Awad, Mohamad M.
    JOURNAL OF APPLIED REMOTE SENSING, 2020, 14 (01):
  • [46] An Evaluation of Popular Hyperspectral Images Classification Approaches
    Kuznetsov, Andrey
    Myasnikov, Vladislav
    EIGHTH INTERNATIONAL CONFERENCE ON MACHINE VISION (ICMV 2015), 2015, 9875
  • [47] CLASSIFICATION OF HYPERSPECTRAL IMAGES BASED ON WEIGHTED DMPs
    Aytekin, Orsan
    Mura, Mauro Dalla
    Ulusoy, Ilkay
    Benediktsson, Jon Atli
    2012 IEEE INTERNATIONAL GEOSCIENCE AND REMOTE SENSING SYMPOSIUM (IGARSS), 2012, : 4154 - 4157
  • [48] Extinction Profiles Fusion for Hyperspectral Images Classification
    Fang, Leyuan
    He, Nanjun
    Li, Shutao
    Ghamisi, Pedram
    Benediktsson, Jon Atli
    IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, 2018, 56 (03): : 1803 - 1815
  • [49] Locality Adaptive Discriminant Analysis for Spectral-Spatial Classification of Hyperspectral Images
    Wang, Qi
    Meng, Zhaotie
    Li, Xuelong
    IEEE GEOSCIENCE AND REMOTE SENSING LETTERS, 2017, 14 (11) : 2077 - 2081
  • [50] Classification of Hyperspectral Images Using Kernel Fully Constrained Least Squares
    Liu, Jianjun
    Wu, Zebin
    Xiao, Zhiyong
    Yang, Jinlong
    ISPRS INTERNATIONAL JOURNAL OF GEO-INFORMATION, 2017, 6 (11)