| 一种干旱半干旱地区决策树分类方法 |
| |
| 引用本文： | 刘清臣,张舞燕,孟秀军.一种干旱半干旱地区决策树分类方法[J].测绘与空间地理信息,2014(7):62-65. |
| |
| 作者姓名： | 刘清臣 张舞燕 孟秀军 |
| |
| 作者单位： | [1]河南理工大学矿山空间信息国家测绘地理信息局重点实验室,河南焦作454000 [2]中国地震局地质研究所地震动力学国家重点实验室,北京100029 |
| |
| 基金项目： | 河南理工大学研究生创新基金（重点）（2011-M-04）资助 |
| |
| 摘 要： | 以内蒙古自治区伊金霍洛旗为研究区,利用Landsat TM影像,对干旱半干旱地区土地利用信息进行提取。在ENVI软件的支持下,分析了影像的光谱特征及NDVI,NDBI,NDWI特征变量,并运用灰度共生矩阵对影像进行纹理特征提取,得到熵纹理特征图像,确定各类地物的阈值,运用决策树分类法对影像进行分类。结果表明,结合光谱特征和纹理特征的决策树分类方法,提取干旱半干旱地区土地利用信息可行且准确性较高。
|
| 关 键 词： | 遥感影像 决策树分类 光谱特征 纹理特征 |
A Classification Method of Decision Tree in Arid and Semi-Arid Regions |
| |
| Institution: | LIU Qing -chen, ZHANG Wu -yan , MENG Xiu - jun ( 1. Henan Polytechnic University, Key Lab. of Mine Spatial Information Technologies of NASMG, Jiaozuo 454000, China; 2. State Key Lab. of Earthquake Dynamics, Institute of Geology, China Earthquake Administration, Beijing 100029, China) |
| |
| Abstract: | Taking case of Ejin Horo County, Inner Mongolia Autonomous Region, the information of land - use is extracled from Landsat TM image in arid and semi-arid regions. With the support of ENVI software, the image is used to analyze the spectral characteris-tics and its value of NDVI, NDBI, and NDWI. The texture infnrmation of image including the entropy is acquired by means of using Gay Level Co - occurrence Matrices. Then the classification results are obtained and the image is classified by the decision tree classi-fication method. The result indicate that the decision tree classification based on spectral and texture features extracting the iniormation of land-use in arid and semi-arid regions has satisfactory performance. |
| |
| Keywords: | remote sensing image decision tree classification spectral features texture features |
| 本文献已被 CNKI 维普 等数据库收录！ |
|
