摘要
采用了辅以纹理特征的决策树方法进行分类,探讨了决策树在遥感数据分类方面的优势,提高了遥感影像的分类精度。
Aiming at the TM images, the decision tree classifier using spectral and texture features is adopted to classification. The experiments prove that the decision tree classifier can get higher accuracy compared with the traditional classification algorism, and can select the features automatically, which decreases the number of the input data.
出处
《测绘信息与工程》
2006年第5期1-3,共3页
Journal of Geomatics
关键词
决策树
分类
纹理
特征选择
TM
遥感
decision tree
classification
texture
feature selection
TM
RS
