基于知识的遥感影像居民地信息提取——以山东省蓬莱市为例  

Residential Information Extraction of RS Image Based on Knowledge——Setting Penglai City of Shandong Province as an Example

作　　者：谭国强[1] 王周龙[1] 王霞[2] 马金卫[1] 

机构地区：[1]鲁东大学地理与规划学院,山东烟台264025 [2]烟台市国土资源局土地规划站,山东烟台264003

出　　处：《山东国土资源》2011年第4期50-53,共4页Shandong Land and Resources

基　　金：山东省自然科学基金重点项目(Z2008E03)

摘　　要：如何自动获得居民地信息是遥感专题信息提取研究的热点问题,该文以山东省蓬莱市为研究区,基于对2004年TM影像光谱特征的分析,建立了基于知识的居民地信息自动提取模型,并对提取结果进行了精度评价。结果表明,利用该模型可以很好地将蓬莱地区的城镇居民地自动提取出来,其提取精度与传统的监督分类方法相比有了很大提高。How to automatically obtain residential information is a key problem in the fields of remote sensing.In this paper,setting Penglai city of Shandong province as a study area,based on spectral analysis of 2004 TM image,automatic information extraction model of residential based on knowledge is set up,and the extraction accuracy of the results have been evaluated.As showed by the results,urban residents in Penglai area can be automatically extracted by using this model,and the extraction accuracy has been greatly improved than traditional method of supervised classification.

关 键 词：TM影像 居民地 信息提取 山东省蓬莱市 

分 类 号：TP75[自动化与计算机技术—检测技术与自动化装置]