摘要
丰富的高分辨率影像及面向对象分类方法为城市土地利用分类与分析提供了新的机遇.以SPOT-5 2.5 m融合后的多光谱遥感影像为主要数据源,采用基于规则的面向对象分类法进行分类.首先分析典型地物的影像光谱特征、几何特征和纹理特征,找出区分点进行,对最终的房屋、工厂、道路、阴影、未利用地5类城市用地进行精度分析,进而对面向对象的分类方法及存在问题进行阐述.
High-spatial-resolution satellite imagery (eg. SPOT) has opened up new opportuni- ties for detailed mapping and analysis of urban land use. Using satellite remote sensing ( SPOT-5 ) , it analysis the detailed of the spectral characteristics , geometric features and textural features in the image , and make full use of them to extract the features . An accuracy assessment was conducted on all the five final land-use classess (building, factory, road, shadow, unused land ). this study will further discuss problems that may be encountered with the object-oriented approach and probable solutions.
出处
《福建师范大学学报(自然科学版)》
CAS
CSCD
北大核心
2014年第6期96-102,共7页
Journal of Fujian Normal University:Natural Science Edition
基金
福建省科技厅产学研重大项目(2012Y4001)