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：Most geographic data sets available today mainly provide two-dimensional geometric information about topographic objects in different aggregation levels. Three dimensional information is distributed separately by means of digital terrain models. In contrast, many applications presume that integrated data sets are available. The extraction of man-made objects from images and laser scanner data has been in the focus of image processing research in recent years. In this paper an approach is presented to identify road objects in laser scanner data sets. The information that is of interest is the outline of the road, its height and further tributary properties like longitudinal and transversal slope, width and curvature. In order to derive the width and further properties of roads, features and attributes of 2D data sets have to be evaluated against their contribution to the integration of height information into a common data model. Planning and construction facilities make use of standardised longitudinal and cross sections. Topographic data sets only contain the middle axis of roads. In this paper the information will be detected, analysed and measured by a model based approach, investigating the neighbourhood of two dimensional road elements in terms of cross sections. After the optimal position of the section is detected, the road parameters can be refined by an estimation procedure. It is subject to further investigations where and how to introduce cross sections into a common data model with respect to handling, storage and quality aspects
关键词：Feature Extraction; Geometric Modelling; High resolution DEM/DTM; Integration
