基于eCognition的城镇绿地信息动态监测研究  

Study on Dynamic Change of Green Land in Town Based on eCognition

作　　者：申广荣[1] 钱振华[1] 徐敬敬[1] 李艳[1] 

机构地区：[1]上海交通大学农业与生物学院,上海200240

出　　处：《上海交通大学学报（农业科学版）》2009年第1期1-6,共6页Journal of Shanghai Jiaotong University(Agricultural Science)

基　　金：2007年上海市科委项目(073119N07);上海市重点学科建设项目(B209)

摘　　要：基于高分辨图像提取绿地信息是分析和掌握城镇绿地分布格局及其动态变化的有效途径。本文借助面向对象的技术,通过对上海奉贤南桥镇不同年份航空遥感图像各类绿地信息的提取研究,分析研究区绿地信息的动态变化。结果表明:基于以面向对象分类技术为核心的eCognition,采用面向对象的多尺度分割方法,并将专题图层直接用于图像的分割和分类,实现不同尺度地物信息的分层提取,对于提取颜色、光谱等特征相似的各类绿地信息,效果良好。更完善的绿地信息分类规则库和技术体系的构建有待于不同区域更多绿地类型的分类实践和相关研究。The extraction of green land information based on the classification of high resolution remote sensing image is efficient way to analyze the dynamic change and distribution pattern of green land in town. This paper, take Nianqiao town of Fenxian district in Shanghai as an example, explores the method and scheme of extracting green land information from the high resolution remote sensing image by using object-oriented technology based on eCognition. The experiments show classification with geometrical feature, multiresolution segmentation and thematic map by eCognition is possible to acquire the distribution patterns of different green land types with high accuracy, which is difficult to distinct by traditional classification scheme since they have similar color and spectral features. The better result would be gained through improving the classification scheme, rules and more practices.

关 键 词：绿地信息 动态监测 ECOGNITION 多尺度分割 专题图 

分 类 号：S127[农业科学—农业基础科学]