基于面向对象的IKONOS影像香榧树分布信息提取研究  

Extraction of Torreya grandis Merrillii based on object-oriented method from IKONOS imagery

作　　者：韩凝[1] 张秀英[1] 王小明[1] 王珂[1] 

机构地区：[1]浙江大学环境与资源学院农业遥感与信息技术应用研究所,浙江杭州310029

出　　处：《浙江大学学报（农业与生命科学版）》2009年第6期670-676,共7页Journal of Zhejiang University:Agriculture and Life Sciences

基　　金：国家自然科学基金资助项目(30671212)

摘　　要：为了定量模拟我国特有经济树种香榧树的生长环境,基于IKONOS卫星影像,采用面向对象的分类方法,充分利用多尺度分割逐级分层提取的优势,对香榧树的分布进行信息提取,以准确地确定香榧树的空间分布.结果表明,其生产者精度为86.36%,用户精度为67.86%,明显高于基于像元的监督分类法的精度.表明基于面向对象的IKONOS影像香榧树分布遥感信息提取方法具有良好的应用价值.In order to model the growth circumstance of Torreya grandis Merrillii, which is a precious economic species peculiar in China, extraction of its distributing information is of great importance. Based on the IKONOS remote sensing imagery, object-oriented method was used, which provided the function of multiresolution segmentation and extraction on different segmented layers gradually. The producer＇s accuracy of identified Torreya grandis Merrillii was 86.360% and the user＇ s accuracy was 67.86%, higher than the result of conventional supervised classification, indicating that the method proposed in this study is feasible to extract Torreya grandis Merrillii information.

关 键 词：香榧 面向对象 信息提取 多尺度分割 

分 类 号：TP751.1[自动化与计算机技术—检测技术与自动化装置]