摘要
采用1994、2001、2008年TM影像作为数据源,运用决策树分类和最小距离分类法,对徐州市城区进行土地利用分类,根据获得的土地利用现状分类图和土地利用转移矩阵,分析徐州市城区的土地利用变化情况,结果表明:(1)决策树法的分类精度要高于最小距离分类法;(2)1994-2001年,耕地迅速减少,变化去向主要是建筑用地,而建筑用地急速增加,主要变化来源是耕地和未利用地;(3)2001-2008年间,耕地继续减少,建筑用地增加,但速度明显放缓。
This paper has taken Xuzhou city as the research object,with the use of remote sensing images of years 1994,2001 and 2008 as data sources.Then through two classification methods,decision tree and minimum distance,we classify the land to five classes.According to the final status of land-use classification maps and Markov transition matrix,we analyse the land-use change of Xuzhou city.The results show that：（1） Decision trees in general,produced consistently higher classification accuracies than traditional minimum distance algorithms.（2） From 1994 to 2001,the farmland has a rapid reduction.The reduced land has mainly changed to urban construction.The land-use of urban construction has increased rapidly.（3）From 2001 to 2008,the farmland is still reduced,but the pace slow down significantly.
出处
《安徽农学通报》
2011年第9期205-207,共3页
Anhui Agricultural Science Bulletin
关键词
遥感
决策树分类
土地利用变化
徐州市城区
Remote sensing
Decision tree classification
Land-use change
Xuzhou city
