文章基本信息

标题：A Supervised Classification Approach Towards Quality Self-Diagnosis of 3D Building Models Using Digital Aerial Imagery
本地全文：下载
作者：Laurence Boudet ; Nicolas Paparoditis ; Franck Jung 等
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：2006
卷号：XXXVI Part 3
出版社：Copernicus Publications
摘要：In the context of 3D building model production or updating, the models have to be manually checked one by one by a human operator in order to ensure their quality. In this paper, we investigate a new approach to perform a quality self-diagnosis of building models in dense urban areas from high resolution aerial images. Hence, we aim at reliably identifying roof facets that do not comply with quality specifications. The self-diagnosis process will highlight potential incorrect facets for their inspection by a human operator. A set of calibrated aerial images enable us to collect positive or negative evidences of roof facet existence and consistency. A particular attention has been paid to the definition of a set of low-level, complementary, robust and consistent image processing measures. Four quality classes have been defined and are used to classify roof facet quality. A supervised classifier and robust decision rules are then applied to perform an effective self-diagnosis according to the traffic light paradigm. Finally, the work in progress leads to a promising quantitative and qualitative evaluation in the context of dense urban areas
关键词：Quality Self-diagnosis; Consistency; Image-based measures; Performance evaluation; Classification; 3D City Models