Interactive 3D Building Modeling and Semantic Labeling from VGI Image Data

Abstract

3D buildings in LoD3 (with façade objects) with semantic information are of great importance for many applications. In this paper, an interactive approach is proposed for 3D building models in LoD3 by using VGI image data. Similar to JSOM, an interactive platform is developed to acquire a façade model with detailed geometries and semantics. Techniques of feature extraction are incorporated in the software, so that a façade can be modeled by only sketching the boundary of the façade and one or two façade objects from one VGI image. The feedbacks of the current Beta version depict that the software is user friendly and can generate 3D façade while labelling façade objects by very few interactive input.