利用决策树分类方法的南京江宁水稻面积变化研究  

Changing of Paddy Area in Jiangning District Based on Decision Tree Classification

作　　者：余超[1] 李明阳[1] 何隆华[2] 

机构地区：[1]南京林业大学,南京210037 [2]中国科学院南京地理与湖泊研究所,南京210008

出　　处：《遥感信息》2015年第2期80-84,共5页Remote Sensing Information

基　　金：国家自然科学基金项目(41271418)

摘　　要：自从2000年12月江宁撤县设为南京市江宁区以来,2002年~2010年期间许多农田现在变成了高楼大厦。该文以南京市江宁区水稻田为研究对象,以江宁区2002年、2006年以及2010年的遥感影像为主要信息源,利用决策树分类的方法,通过坡度、湿度、NDVI等决策因子,根据野外实地勘测调查数据对比,提取出江宁区的水稻田,再通过驱动因子分析,用地理加权回归方法找出主要影响因子,加上空间分布格局的动态分析,从而比较出在这段时间内南京江宁区水稻田的变化情况。研究表明:1江宁区在2002年~2010年内水稻田面积减少了约690km2,约70%的稻田变为其他用地,变化非常明显;2主要影响面积变化的驱动因子是人类居住活动范围的变化;3水稻田的空间格局在2002年~2010年间聚集性减弱,破碎化趋势加强。Significant change has taken place in Jiangning from 2002 to 2010.Since Jiangning became one of the districts of Nanjing in December 2000,much farmland has turned into high-rise buildings.This paper takes the paddy rice in Jiangning district as the research topic.It uses the remote sensing images of Jiangning district in 2002,2005 and 2011 as the main source of information.Adopting a decision tree classification method with slope,humidity,NDVI and other decision-making as factors,this paper extracts the rice fields in Jiangning district according to survey data compared with field surveys.The change area of rice fields in Jiangning in the past decade is obtained by driving factor analysis,using geographically weighted regression to identify the main factor,coupled with the dynamic analysis of spatial distribution pattern.Studies have shown that:(1)The area of rice fields in Jiangning district during 2002 to 2010 reduced by about 690km2 .With about 70% of the paddy fields turned into other uses,change is very obvious.(2 )The main driving factor is the change in the scope of human habitation activities.(3 )The aggregation of spatial pattern in the rice fields between 2002 and 2010 has decreased,and the fragmentation trend has strengthened.

关 键 词：水稻 决策树 地理加权回归 空间格局 面积变化 地理国情监测 

分 类 号：S511[农业科学—作物学] TP79[自动化与计算机技术—检测技术与自动化装置]