Auto-classification of Landform in Karst Region Based on ASTER GDEM
Received date: 2010-11-10
Request revised date: 2011-01-18
Online published: 2012-03-25
Copyright
Geomorphology is one of the most important parts which constitute the elements of physical geography. Based on the GDEM of 1:100000 ASTER, the optimum analysis window was verified and topographic factors were extracted in the unit of drainage area. With GIS spatial analysis and unsupervised classification, the general geomorphologic types in Karst Region were auto-classified. The study results indicate: (1) DEM at the scale of 1:100000 can fill the requirements to express the configuration of earth surface on meso-scale. (2) It confirms the basic regulation to select the analysis window and extract topographic factors taking the drainage area as a unit. Topographic factors extracted can reflect the actual configuration of earth surface more impersonally. (3) Multi-spectral image is combined with topographic factors. With the method of ISODATA unsupervised classification, it can implement the quantification of the general geomorphologic types and automatic classification effectively on meso-scale. The precision of the data extracted can meet the demands of computer automatic classification. These experimental results improve the application of ISODATA unsupervised classification in the automatic classification of geomorphology.
Key words: geomorphic type; ASTER GDEM; spatial analysis; auto-classification
MA Shi-bin , AN Yu-lun . Auto-classification of Landform in Karst Region Based on ASTER GDEM[J]. SCIENTIA GEOGRAPHICA SINICA, 2012 , 32(3) : 368 -373 . DOI: 10.13249/j.cnki.sgs.2012.03.368
Fig.1 The sketch map of the location of study area图1 研究区区域位置 |
Fig.2 The ASTER GDEM data and watershed division of study area图2 研究区ASTER GDEM数据和流域划分 |
Table 1 The best analysis window size of each basin表 1 各流域最佳分析窗口大小 |
流域号 | 最佳分析窗口大小 (栅格数) | DEM标准差 |
---|---|---|
1 | 50×50 | 277.99 |
2 | 40×40 | 223.77 |
3 | 40×40 | 208.61 |
4 | 46×46 | 238.59 |
5 | 46×46 | 241.91 |
6 | 40×40 | 208.72 |
7 | 46×46 | 241.29 |
8 | 50×50 | 260.8 |
Table 2 The classification system of geomorphological type表2 地貌基本形态类型分类 |
地貌分类 | 次级地貌分类 | 分类指标 | 注 释 | |
---|---|---|---|---|
相对高度(m) | 绝对高度(m) | |||
盆(坝)地 | 低盆(坝)地 | 0~900 | 盆底坡度<5° | |
中盆(坝)地 | <30 | 900~1900 | ||
高盆(坝)地 | >1900 | |||
丘陵 | 低 丘 | >900 | ||
中 丘 | 30~200 | 900~1900 | ||
高 丘 | >1900 | |||
低山 | - | >200 | 0~900 | |
中山 | 低中山 | 900~1400 | ||
中中山 | >200 | 1400~1900 | ||
高中山 | >1900 |
Table 3 Correlation between the various landform ingredients表3 各地形因子间相关系数 |
高程 | 坡度 | 全累计变率 | 坡度变率 | DEM晕渲图 | 地形起伏度 | 地表切割 深度 | 地表 粗糙度 | 高程变异系数 | |
---|---|---|---|---|---|---|---|---|---|
高程 | 1.000 | ||||||||
坡度 | -0.136 | 1.000 | |||||||
全累计变率 | 0.207 | 0.088 | 1.000 | ||||||
坡度变率 | -0.203 | 0.247 | -0.012 | 1.000 | |||||
DEM晕渲图 | 0.048 | 0.005 | -0.013 | -0.108 | 1.000 | ||||
地形起伏度 | -0.385 | 0.476 | 0.011 | 0.180 | -0.110 | 1.000 | |||
地表切割深度 | -0.202 | 0.329 | 0.019 | 0.244 | -0.079 | 0.901 | 1.000 | ||
地表粗糙度 | -0.273 | 0.931 | 0.057 | 0.490 | -0.194 | 0.417 | 0.296 | 1.000 | |
高程变异系数 | -0.557 | 0.374 | -0.001 | 0.311 | -0.124 | 0.885 | 0.640 | 0.320 | 1.000 |
Fig.3 The topographic factors derived from ASTER GDEM of the study area图3 DEM 提取的研究区地形因子 |
Fig.4 The morphometric classification map of the study area图4 研究区地貌分类 |
Table 4 Error matrix of classification signature表4 误差矩阵 |
中盆(坝)地 | 高盆(坝)地 | 中丘 | 高丘 | 低中山 | 中山 | 高中山 | 总和 | |
---|---|---|---|---|---|---|---|---|
中盆(坝)地 | 16 | 0 | 13 | 0 | 4 | 0 | 0 | 33 |
高盆(坝)地 | 0 | 25 | 0 | 6 | 0 | 0 | 0 | 31 |
中丘 | 3 | 0 | 28 | 0 | 6 | 6 | 0 | 43 |
高丘 | 4 | 6 | 0 | 24 | 2 | 1 | 5 | 42 |
低中山 | 1 | 0 | 0 | 0 | 41 | 3 | 2 | 47 |
中山 | 3 | 1 | 3 | 2 | 2 | 73 | 3 | 87 |
高中山 | 0 | 1 | 1 | 0 | 2 | 5 | 48 | 57 |
总精度=75% |
Table 5 Precision evaluation of the classification result表5 精度评价(%) |
制图精度 | 漏分误差 | 用户精度 | 错分误差 | |||||
---|---|---|---|---|---|---|---|---|
ASTERGDEM | srtm-DEM | ASTERGDEM | srtm-DEM | ASTERGDEM | srtm-DEM | ASTERGDEM | srtm-DEM | |
中盆(坝)地 | 39.39 | 37.5 | 60.61 | 62.5 | 54.17 | 30 | 45.83 | 70 |
高盆(坝)地 | 48.39 | 80 | 51.61 | 20 | 65.22 | 40 | 34.78 | 60 |
中丘 | 65.12 | 65.12 | 34.88 | 34.88 | 59.57 | 70 | 40.43 | 30 |
高丘 | 57.14 | 45.24 | 42.86 | 54.76 | 61.54 | 82.61 | 38.46 | 17.39 |
低中山 | 87.23 | 82.98 | 12.77 | 17.02 | 71.93 | 76.47 | 28.07 | 30.77 |
中山 | 83.91 | 80.46 | 16.09 | 19.54 | 80.22 | 76.92 | 19.78 | 23.08 |
高中山 | 84.21 | 78.95 | 15.79 | 21.05 | 82.76 | 70.31 | 17.24 | 29.69 |
注:表中srtm-DEM提取地貌类型数据精度参数来自作者硕士毕业论文。 |
The authors have declared that no competing interests exist.
[1] |
|
[2] |
|
[3] |
|
[4] |
|
[5] |
|
[6] |
|
[7] |
|
[8] |
|
[9] |
|
[10] |
贵州省农业区划办编.贵州省农业地貌区划[M].贵阳:贵州人民出版社,1989.
|
[11] |
|
[12] |
|
[13] |
|
[14] |
|
[15] |
|
[16] |
|
[17] |
|
[18] |
|
[19] |
|
[20] |
|
[21] |
|
[22] |
|
[23] |
|
[24] |
|
[25] |
|
/
〈 | 〉 |