摘要
采用面向对象影像分析方法,以成都郫县古城镇土地利用信息提取为研究内容,对预处理后的QuickBird影像进行多层次分割,建立分类体系,定义相应知识库,综合运用模糊分类器和最邻近分类器进行土地利用信息提取研究。试验区影像提取OA为87.04%,Kappa系数为0.8434,结果表明,面向对象影像分析能准确、快速的获取土地利用信息。
The extraction of the land information in Gucheng , Pixian, Chengdu was studied by the object-oriented image analysis. The pretreatment QuickBird image was segmented in multi-level, the classification system was established and the corresponding knowledge base was defined. The extraction of the land application information was analyzed by the fuzzy classifier and the nearest-classifier. The image extraction OA of the experimental area was 87.04 %, Kappa coefficient was 0. 8434. The results showed that the land-use information could be accurately and quickly extracted by the object-oriented image analysis .
出处
《西南农业学报》
CSCD
北大核心
2009年第3期733-738,共6页
Southwest China Journal of Agricultural Sciences
基金
国家"十一五"科技支撑计划(2006BAJ05A13)
关键词
面向对象
信息
提取
Object-oriented
Information
Extraction
