基于遥感数据SOFM网络分类的五种城市增长方式鉴别方法应用研究

¹ 中国科学院遥感应用研究所,北京市朝阳区大屯路9718信箱, 100101²北京师范大学地理与遥感学院,北京市新街口外大街19号,100875;³中国地质大学国土资源与高新技术研究中心,北京市海淀区学院路29号,100083

基金项目: 国家十五重点科技攻关计划资助项目(200BA9048072);国家863计划资助项目(2003AA1350802)

详细信息

Analysis of Five Forms of Urban Expansion Identification Model Based on SOFM Classification Results of Remote Sensing Data

¹ Institute of Remote Sensing Applications, CAS, 9718 Box, Datun Road, Beijing 100101, China; ² Department of Geography, Beijing Normal University, 19 Xinjiekouwai Street, Beijing 100875, China; ³ Institute of Land Resources and High Techniques, China University of Geosciences, 29 Xueyuan Road, Beijing 100083, China;

Funds: 国家十五重点科技攻关计划资助项目(200BA9048072);国家863计划资助项目(2003AA1350802)

摘要: 使用了自主研发的自组织神经网络分类(SOFM)方法[1～3],选择了1988、1994、2001和2003年5～6月份TM+时间序列多光谱遥感数据,对北京城市增长方式进行了30m分辨率遥感时序数据的鉴别,包括填充式增长方式、扩张式增长方式、独立式增长方式、线状式增长方式和簇状增长方式,并绘制了三个时期的城市增长图。在此基础上,根据北京城市增长环线驱动的特点,分别对四环内、四环至五环、五环至六环1988～1994年、1994～2001年、2001～2003年的5种城市扩展方式面积进行了统计。
- 神经网络分类 /
- 城市五种增长方式鉴别模型 /
- 城市环线增长面积统计
Abstract: This paper applies the SOFM (self organized feature map of neural network) classification software, and choose the Landsat temporal TM images in Mays or Junes of 1988, 1994, 2001 and 2003 and applies the five forms of urban expansion identification model considering the ring roads. The results shows that infillment and expansion are the main forms in the three periods of urban growth process from 1988 to 2001, from 1994 to 2001 and from 2001 to 2003, while the linear branch and clustered branch are the main forms in the urban growth from 1994 to 2001 between the fifth and sixth ring roads.
- neural network classification /
- five forms of urban expansion identification model /
- statistics of urban ring expansion area