• Journal of Geo-information Science
  • Vol. 22, Issue 10, 2088 (2020)
Jie YE1, Fanxiao MENG1,*, Weiming BAI1, Bin ZHANG1, and Jinming ZHENG2
Author Affiliations
  • 1Henan Aero Geophysical Survey and Remote Sensing Center, Zhengzhou 450053, China
  • 2Northwest Institute of Nuclear Technology, Xi'an 710024, China
  • show less
    DOI: 10.12082/dqxxkx.2020.190483 Cite this Article
    Jie YE, Fanxiao MENG, Weiming BAI, Bin ZHANG, Jinming ZHENG. A Comparative Study on the Classification of GF-1 Remote Sensing Images for Zhoukou Urban under the Four Identical Condition[J]. Journal of Geo-information Science, 2020, 22(10): 2088 Copy Citation Text show less
    Fusion remote sensing image of GF-1 satellite for the main districts of Zhoukou acquired on April 17, 2018
    Fig. 1. Fusion remote sensing image of GF-1 satellite for the main districts of Zhoukou acquired on April 17, 2018
    Workflow of object-based classification in the study area
    Fig. 2. Workflow of object-based classification in the study area
    The distribution of training, verification sample and investigation site
    Fig. 3. The distribution of training, verification sample and investigation site
    Comparison of pixel-based and object-based classification under three machine learning classifiers including CART, SVM and RF using GF-1 remote sensing image for classing the main district of Zhoukou urban
    Fig. 4. Comparison of pixel-based and object-based classification under three machine learning classifiers including CART, SVM and RF using GF-1 remote sensing image for classing the main district of Zhoukou urban
    Comparison of producer's accuracy and user's accuracy of pixel-based classification(under three machine learning classifiers including CART, SVM and RF) for Zhoukou urban at class level
    Fig. 5. Comparison of producer's accuracy and user's accuracy of pixel-based classification(under three machine learning classifiers including CART, SVM and RF) for Zhoukou urban at class level
    Comparison of producer's accuracy and user's accuracy of object-based classification(under three machine learning classifiers including CART, SVM and RF) for Zhoukou urban at class level
    Fig. 6. Comparison of producer's accuracy and user's accuracy of object-based classification(under three machine learning classifiers including CART, SVM and RF) for Zhoukou urban at class level
    Comparison of local details of pixel-based and object-based classification for Zhoukou urban under three machine learning classifiers including CART, SVM and RF
    Fig. 7. Comparison of local details of pixel-based and object-based classification for Zhoukou urban under three machine learning classifiers including CART, SVM and RF
    参数2 m分辨率全色/8 m分辨率多光谱16 m分辨率多光谱
    光谱范围/μm全色0.45~0.90-
    多光谱0.45~0.520.45~0.52
    0.52~0.590.52~0.59
    0.63~0.690.63~0.69
    0.77~0.890.77~0.89
    空间分辨率/m全色216
    多光谱8
    幅宽/km60(2台相机)800(4台相机)
    Table 1. The parameters of GF-1 satellite image
    特征类型特征名称物理意义
    光谱特征Mean RR波段均值
    Ratio RR波段比率
    quantile[50] (R)R波段分位数
    Max.diff.最大差值
    Standard deviation NIRNIR波段标准偏差
    HIS Transformation SaturationHIS空间饱和度
    形状特征AreaShape indexDensityElliptic FitCompactness(polygon)面积
    形状指数
    密度
    椭圆拟合率
    紧致度
    纹理特征GLCM Entropy(all.dir.)GLCM Homogeneity(all.dir.)灰度共生矩阵熵
    灰度共生矩阵均质性
    自定义特征NDWINDVI归一化水体指数
    归一化植被指数
    Table 2. Optimal feature set for object-based classifications
    面向像元CART分类混淆矩阵面向对象CART分类混淆矩阵
    农业用地林草地水体湿地建筑用地交通用地总计UA/%农业用地林草地水体湿地建筑用地交通用地总计UA/%
    农业用地6460948377.10农业用地7130017594.67
    林草地4590537183.10林草地2610106495.31
    水体湿地0047805585.45水体湿地1050005198.04
    建筑用地1204513310051.00建筑用地21180109485.10
    交通用地00016395570.90交通用地4008688085.00
    总计8065518979总计8065518979
    PA/%80.0090.7792.1657.3049.37PA/%88.7593.8598.0489.8986.08
    OA=71.43% Kappa=0.64OA=90.66% Kappa=0.88
    面向像元SVM分类混淆矩阵面向对象SVM分类混淆矩阵
    农业用地林草地水体湿地建筑用地交通用地总计UA/%农业用地林草地水体湿地建筑用地交通用地总计UA/%
    农业用地7120438088.75农业用地7210137793.50
    林草地3630306991.30林草地1640316992.75
    水体湿地00480048100.00水体湿地0051105298.08
    建筑用地503603510358.25建筑用地5007948889.77
    交通用地10022416464.06交通用地2005717891.03
    总计8065518979总计8065518979
    PA/%88.7596.9294.1167.4251.90PA/%90.0098.46100.0088.7689.87
    OA=77.75% Kappa=0.71OA=92.58% Kappa=0.91
    面向像元RF分类混淆矩阵面向对象RF分类混淆矩阵
    农业用地林草地水体湿地建筑用地交通用地总计UA/%农业用地林草地水体湿地建筑用地交通用地总计UA/%
    农业用地6630958379.52农业用地7210017497.30
    林草地7620637879.49林草地1640306894.12
    水体湿地0049305294.23水体湿地00510051100.00
    建筑用地60250147269.44建筑用地5008279487.23
    交通用地10021577972.15交通用地2004717792.20
    总计8065518979总计8065518979
    PA/%82.5095.3896.0856.1872.15PA/%90.0098.46100.0092.1389.87
    OA=78.02% Kappa=0.72OA=93.40% Kappa=0.92
    Table 3. Confusion matrices and associated classifier accuracies based on pixel-based and object-based classifications under three machine learning classifiers including CART, SVM and RF using GF-1 remote sensing image for classing the main district of Zhoukou urban
    Jie YE, Fanxiao MENG, Weiming BAI, Bin ZHANG, Jinming ZHENG. A Comparative Study on the Classification of GF-1 Remote Sensing Images for Zhoukou Urban under the Four Identical Condition[J]. Journal of Geo-information Science, 2020, 22(10): 2088
    Download Citation