Remote Sensing Technology and Application, Volume. 39, Issue 3, 708(2024)

Extraction Method of Alpine Wetland Information Using Landsat Data

Lu CHEN, Wangping LI, Junming HAO, Zhaoye ZHOU, Xiuxia ZHANG, Xiaoqiang CHENG, and Xiaoxian WANG
Author Affiliations
  • School of Civil Engineering, Lanzhou University of Technology, Lanzhou730050,China
  • show less
    Figures & Tables(11)
    Geographical Location Diagram of Luqu County
    Distribution diagram of training sample points and validation sample points in the study area
    Percentage contribution rate and displacement importance of "25 features" calculated by Jackknife tool
    Correlation coefficient matrix of 25 feature factors obtained by ArcGIS software
    Contribution rate and replacement importance of optimal features collection of different wetland types
    Wetland distribution map of Luqu County
    • Table 1. Wetland classification feature set

      View table
      View in Article

      Table 1. Wetland classification feature set

      特征变量指数指数名称特征因子说明/公式文献
      光谱特征B波段B1~B7,共7个波段[18]
      Y1缨帽变换亮度分量反映总体反射值[19]
      Y2缨帽变换绿度分量反映绿色生物量特征[19]
      Y3缨帽变换湿度分量反映湿度特征[19]
      植被、水体、土地利用指数CIWI混合水体指数CIWI=NDVI+NIR+C[20]
      NDBI归一化土地利用指数NDBI=(SWIR1-NIR)/(SWIR1+NIR)[21]
      NDVI归一化植被指数NDVI=(NIR-Red)/(NIR+Red)[22]
      NDWI归一化差异水体指数NDWI=(Green-NIR)/(Green+NIR)[22]
      NDWI_BNDWI升级版NDWI=(Green-NIR)/(Blue+NIR)[23]
      地形特征DEM数字高程模型反映垂直高度信息[23]
      Slop坡度Slop=(h/l)*100%[24]
      Aspect坡向法线的正方向在水平面上的投影与正北方向的夹角[15]
      Tip地形位置指数TPI=Z—Z¯[25]
      纹理特征Mean均值Mean=1MNI=0m-1j=0N-1f(i,j,d,θ)[25]
      Var方差Var=I=0m-1j=0N-1(i-μ)2f(i,j,d,θ)[25]
      EntorpyEnt=-I=0m-1j=0N-1f(i,j)lgf(i,j,d,θ)[26]
      Asm角二阶矩ASM=I=0m-1j=0N-1f(i,j,d,θ)[26]
      Dissimilarity相异性Dis=I=0m-1j=0N-1i-jf(i,j,d,θ)[26]
      Homoeneity同质性Homo=I=0m-1j=0N-1f(i,j,d,θ)1+(i-j)2)[26]
    • Table 2. <i>C</i><sub>1</sub>、<i>C</i><sub>2</sub>、<i>C</i><sub>3</sub> values table

      View table
      View in Article

      Table 2. <i>C</i><sub>1</sub>、<i>C</i><sub>2</sub>、<i>C</i><sub>3</sub> values table

      类型判断条件相似程度C1C2C3
      类型1μ/σ>20湿地或非湿地0.90.71.0
      类型23<μ/σ≤20湿地边界1.01.01.0
      类型3μ/σ≤3,μ≤0.25小面积湿地0.71.12.0
      类型4μ/σ≤3,μ>0,25其他0.60.81.0
    • Table 3. Optimal classification feature set of each wetland type

      View table
      View in Article

      Table 3. Optimal classification feature set of each wetland type

      湿地类型最优特征因子集
      河流湿地Var、Nir、K-T Y3、K-T Y2、DEM
      湖泊湿地Nir、DEM、Mean、Var
      沼泽湿地Slop、DEM、Green、 Swir2、NDWI_B
      冰川湿地Blue、DEM、Red、K-T Y1
    • Table 4. Prediction results of “25 Features” and preferred features AUC value of area under ROC curve

      View table
      View in Article

      Table 4. Prediction results of “25 Features” and preferred features AUC value of area under ROC curve

      训练集AUC值测试集AUC值随机预测AUC值
      “25特征”优选特征“25特征”优选特征“25特征”优选特征
      河流湿地0.9780.9740.9770.9730.50.5
      湖泊湿地0.9880.9880.9860.9890.50.5
      沼泽湿地0.8960.9460.8550.9490.50.5
      冰川湿地0.9950.9930.9940.9920.50.5
    • Table 5. Statistical table of wetland classification accuracy

      View table
      View in Article

      Table 5. Statistical table of wetland classification accuracy

      分类后

      数据/个

      参考数据总和
      其他河流湿地湖泊湿地沼泽湿地冰川湿地
      其他209452821267
      河流湿地17303077
      湖泊湿地00465051
      沼泽湿地3051290137
      冰川湿地00007171
      总和213775616592432
      制图精度98.12%94.81%82.14%78.18%77.17%
      用户精度78.28%94.81%90.20%94.16%100%
    Tools

    Get Citation

    Copy Citation Text

    Lu CHEN, Wangping LI, Junming HAO, Zhaoye ZHOU, Xiuxia ZHANG, Xiaoqiang CHENG, Xiaoxian WANG. Extraction Method of Alpine Wetland Information Using Landsat Data[J]. Remote Sensing Technology and Application, 2024, 39(3): 708

    Download Citation

    EndNote(RIS)BibTexPlain Text
    Save article for my favorites
    Paper Information

    Category:

    Received: Oct. 13, 2022

    Accepted: --

    Published Online: Dec. 9, 2024

    The Author Email:

    DOI:10.11873/j.issn.1004-0323.2024.3.0708

    Topics