• 北大核心期刊(《中文核心期刊要目总览》2017版)
  • 中国科技核心期刊(中国科技论文统计源期刊)
  • JST 日本科学技术振兴机构数据库(日)收录期刊

留言板

尊敬的读者、作者、审稿人, 关于本刊的投稿、审稿、编辑和出版的任何问题, 您可以本页添加留言。我们将尽快给您答复。谢谢您的支持!

姓名
邮箱
手机号码
标题
留言内容
验证码

空谱协同竞争保持图嵌入高光谱图像特征提取

陶洋 翁善 林飞鹏 杨雯

陶洋, 翁善, 林飞鹏, 杨雯. 空谱协同竞争保持图嵌入高光谱图像特征提取[J]. 微电子学与计算机, 2021, 38(9): 17-22.
引用本文: 陶洋, 翁善, 林飞鹏, 杨雯. 空谱协同竞争保持图嵌入高光谱图像特征提取[J]. 微电子学与计算机, 2021, 38(9): 17-22.
TAO Yang, WENG Shan, LIN Feipeng, YANG Wen. Feature extraction of hyperspectral image with spatial-spectral collaboration-competition preserving graph embedding[J]. Microelectronics & Computer, 2021, 38(9): 17-22.
Citation: TAO Yang, WENG Shan, LIN Feipeng, YANG Wen. Feature extraction of hyperspectral image with spatial-spectral collaboration-competition preserving graph embedding[J]. Microelectronics & Computer, 2021, 38(9): 17-22.

空谱协同竞争保持图嵌入高光谱图像特征提取

详细信息
    作者简介:

    陶洋  男,(1964-),博士研究生,教授.研究方向为异构网络、图像处理等.E-mail: 1cqtxrjyjs@126.com

    翁善  男,(1991-),硕士研究生.研究方向为图像处理

    林飞鹏  男,(1992-),硕士研究生.研究方向为图像处理

    杨雯  女,(1992-),硕士研究生.研究方向为图像处理

  • 中图分类号: TP751

Feature extraction of hyperspectral image with spatial-spectral collaboration-competition preserving graph embedding

  • 摘要: 针对高光谱数据样本标签标注困难问题,以及多数特征提取算法仅考虑光谱特征信息,而忽略了空间信息的问题.提出了一种在无监督场景下的空谱协同竞争保持图嵌入(SCPGE)高光谱图像特征提取方法.利用协同表示揭示全局流形结构,配合基于空间近邻信息和光谱近邻信息的局部约束特性来计算出该像元的表示系数,继而利用表示系数矩阵构建图的权重矩阵,通过施加正则项的图嵌入目标函数获得最佳投影矩阵.在公开的数据集Indian Pines和Salina数据集上验证表明,所提算法与其它同类算法具有较优的结果.
  • 图  1  SCPGE算法流程图

    图  2  SCPGE在不同参数下的OA

    图  3  SCPGE在不同特征提取维数d下的OA

    图  4  SCPGE在不同训练样本占比下的OA

    表  1  Indian Pines数据集样本情况及分类结果

    类别 训练集 测试集 LPP NPE SPGE SLPGE CPGE CCPGE SCPGE
    Alfalfa 5 41 22.93 9.51 13.66 35.12 39.02 37.07 52.93
    Corn-notill 143 1285 60.65 59.84 62.12 61.93 68.51 69.49 78.83
    Corn-mintill 83 747 46.25 35.77 51.55 51.87 60.11 61.37 73.52
    Corn 24 213 15.4 13.94 38.4 32.63 41.27 41.31 61.6
    Grass-pasture 48 435 78.16 81.61 85.43 87.54 89.13 88.9 91.31
    Grass-trees 73 657 89.79 92.4 92.36 94.26 95.19 94.05 95.86
    Grass-pasture-mowed 3 25 38.8 26 60 48.8 47.2 54.8 50.4
    Hay-windrowed 48 430 98.21 98.21 98.19 98.7 98.4 98.35 98.84
    Oats 3 17 13.53 7.65 24.12 21.18 18.24 28.82 46.47
    Soybean-notill 97 875 44.7 43.74 50.99 51.23 58.65 60.62 72.66
    Soybean-mintill 246 2209 80.32 81.01 77.58 81.01 83.5 82.74 83.68
    Soybean-clean 59 534 27.73 24.81 44.49 51.7 55.54 58.73 79.76
    Wheat 21 184 89.89 90.54 91.25 91.58 94.18 94.24 97.72
    Woods 127 1138 93.78 94.43 92.71 95.07 95.74 94.67 95.43
    Buildings-Grass-Trees-Drives 39 347 43.37 42.1 49.63 44.64 48.16 49.68 62.77
    Stone-Steel-Towers 9 84 70.48 72.38 67.74 81.79 81.9 77.26 80.12
    OA 68.18 67.4 70.93 72.61 76.32 76.57 82.95
    AA 57.12 54.62 62.51 64.32 67.17 68.26 76.37
    Kappa 63.01 62.06 66.48 68.35 72.71 73.05 80.48
    下载: 导出CSV

    表  2  Salinas数据集样本情况及分类结果

    类别 训练集 测试集 LPP NPE SPGE SLPGE CPGE CCPGE SCPGE
    Brocoli_green_weeds_1 100 1909 99.13 99.2 99.57 99.5 99.43 99.33 99.46
    Brocoli_green_weeds_2 186 3540 99.86 99.82 99.85 99.82 99.84 99.9 99.78
    Fallow 99 1877 98.15 98.72 98.6 99.25 97.72 99.26 99.65
    Fallow_rough_plow 70 1324 96.88 96.47 98.8 98.87 98.92 98.94 99.26
    Fallow_smooth 134 2544 98.12 98.64 99.02 99.26 98.85 99.28 99.39
    Stubble 198 3761 99.84 99.83 99.76 99.74 99.77 99.77 99.75
    Celery 179 3400 99.88 99.88 99.82 99.82 99.86 99.89 99.82
    Grapes_untrained 564 10707 87.94 87.21 88.52 89.59 89.25 89.04 90.54
    Soil_vinyard_develop 310 5893 99.61 99.58 99.63 99.8 99.81 99.75 99.82
    Corn_senesced_green_weeds 164 3114 94.7 95.9 96.75 97.13 96.53 96.76 97.47
    Lettuce_romaine_4wk 53 1015 96.3 95.74 97.8 98.21 97.6 97.41 97.85
    Lettuce_romaine_5wk 96 1831 97.74 99.18 99.34 99.83 99.97 99.86 99.97
    Lettuce_romaine_6wk 46 870 93.41 95.77 98.84 99.13 99.41 99.15 99.14
    Lettuce_romaine_7wk 54 1016 92.84 94.74 96.63 97.46 96.59 96.53 97.19
    Vinyard_untrained 363 6905 67.55 68.56 70.44 72.46 74.26 74.04 77.57
    Vinyard_vertical_trellis 90 1717 98.96 99.07 99.28 99.33 99.16 99.15 99.34
    OA 92.02 92.23 93.04 93.65 93.69 93.69 94.58
    AA 95.06 95.52 96.42 96.82 96.69 96.75 97.25
    Kappa 91.1 91.34 92.24 92.92 92.97 92.97 93.96
    下载: 导出CSV
  • [1] LI W, WANG Z J, LI L, et al. Feature extraction for hyperspectral images using local contain profile[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2019, 12(12): 5035-5046. DOI:  10.1109/JSTARS.2019.2951437.
    [2] 黄鸿, 唐玉枭, 段宇乐. 半监督多图嵌入的高光谱影像特征提取[J]. 光学精密工程, 2020, 28(2): 443-456. DOI:  10.3788/OPE.20202802.0443.

    HUANG H, TANG Y X, DUAN Y L. Feature extraction of hyperspectral image with semi-supervised multi-graph embedding[J]. Optics and Precision Engineering, 2020, 28(2): 443-456. DOI:  10.3788/OPE.20202802.0443.
    [3] ZHANG L P, LUO F L. Review on graph learning for dimensionality reduction of hyperspectral image[J]. Geo-spatial Information Science, 2020, 23(1): 98-106. DOI:  10.1080/10095020.2020.1720529.
    [4] LY N H, DU Q, FOWLER J E. Sparse graph-based discriminant analysis for hyperspectral imagery[J]. IEEE Transactions on Geoscience and Remote Sensing, 2014, 52(7): 3872-3884. DOI:  10.1109/TGRS.2013.2277251.
    [5] LI W, LIU J B, DU Q. Sparse and low-rank graph for discriminant analysis of hyperspectral imagery[J]. IEEE Transactions on Geoscience and Remote Sensing, 2016, 54(7): 4094-4105. DOI:  10.1109/TGRS.2016.2536685.
    [6] HUANG S, YU Y, YANG D, et al. Collaborative graph embedding: a simple way to generally enhance subspace learning algorithms[J]. IEEE Transactions on Circuits and Systems for Video Technology, 2015, 26(10): 1835-1845. DOI:  10.1109/TCSVT.2015.2455751.
    [7] LIU N, LI W, DU Q. Unsupervised feature extraction for hyperspectral imagery using collaboration-competition graph[J]. IEEE Journal of Selected Topics in Signal Processing, 2018, 12(6): 1491-1503. DOI:  10.1109/JSTSP.2018.2877474.
    [8] LUO F L, GUO T, LIN Z P, et al. Semisupervised hypergraph discriminant learning for dimensionality reduction of hyperspectral image[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2020(13): 4242-4256. DOI:  10.1109/JSTARS.2020.3011431.
    [9] 黄鸿, 陈美利, 段宇乐, 等. 空-谱协同流形重构的高光谱影像分类[J]. 光学精密工程, 2018, 26(7): 1827-1836. DOI:  10.3788/OPE.20182607.1827.

    HUANG H, CHEN M L, DUAN Y L, el al. Hyper-spectral image classification using spatial-spectral manifold reconstruction[J]. Optics and Precision Engineering, 2018, 26(7): 1827-1836. DOI:  10.3788/OPE.20182607.1827.
  • 加载中
图(4) / 表(2)
计量
  • 文章访问数:  45
  • HTML全文浏览量:  26
  • PDF下载量:  9
  • 被引次数: 0
出版历程
  • 收稿日期:  2021-02-01
  • 修回日期:  2021-03-03
  • 刊出日期:  2021-09-05

目录

    /

    返回文章
    返回