Ninth International Conference on The Constructed Environment

Abstract

The implementation of building information model (BIM) enhances facility management. It aims to extend the building information application from 2.5 dimensions to 5 dimensions which enables the uses of spatial data with temporal and costing dimension. However, most of the constructed environment are lack of existing building information. Surveyors need to have intensive measurement for building information for better facility management. With the development of laser scanning technology, the massive building information would be collected with safe, reliable, and effective approaches. Researchers have lots of attempts to develop effective and efficient approaches for building objects extraction from unorganized point cloud. This paper proposes a new approach to extract and recognize the building object information with the considering of the object nature, building style and the use machine learning. The result of this study is presented and evaluated.