文章基本信息

标题：GENERALIZATION OF BUILDING POLYGONS EXTRACTED FROM IKONOS IMAGERY
本地全文：下载
作者：Jie Shan ; D.Scott Lee
期刊名称：ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences
印刷版ISSN：2194-9042
电子版ISSN：2194-9050
出版年度：2002
卷号：XXXIV Part 4
出版社：Copernicus Publications
摘要：Building extraction in this paper is treated as a classification guided image segmentation problem followed by a generalization operation. Initial building locations are first found by using supervised classification on IKONOS multispectral images. A supervised classification/cluster based approach is then used to segment the building image from its background and surroundings in the panchromatic image. A series of polygon operations are implemented to determine the best delineation of buildings based on the segmentation results. The delineated building polygons are subsequently refined by a two-step generalization process, distance-based generalization and curvature-based generalization. The distance-based generalization uses the classical Douglas-Peucker approach to remove redundant nodes in the building polygons, while a curvature-based approach is proposed to eliminate sharp and unrealistic angles in the building polygons. Results from IKONOS imagery are presented to show the implementation and efficiency of the proposed methodology
关键词：Building extraction; Generalization; Shaping; Satellite Imagery; High resolution