Answers are grounded in this paper and cite their sources.
In early stages of fast tracked construction projects—such as mega industrial projects—contractors do not have access to detailed and final drawings. Yet, they usually need rough estimates of material quantities in the project. They usually depend on initial versions of BIM models that contain all trades (e.g., pipes, steel, etc.) information. Because they are premature models, 3D object attributes may be missed, omitted, inconsistent, or incomplete which makes extracting quantities from them a manual, time-consuming, and inaccurate task. In this research, we aim to automate this task by applying shape recognition techniques on unlabeled 3D models. These techniques have been used in different fields such as handwriting recognition, object recognition, and 3D object search and retrieval. We experiment with some of these techniques to estimate material quantities for industrial construction projects based on the geometry of unlabeled 3D objects. We investigate a number of these techniques keeping in mind that the main usage will be a preliminary estimate in early stages of the project; therefore, we favor fast techniques with a margin of error over more accurate—but relatively slow—ones. This paper discusses applicability of these techniques and shows the results of applying and testing them on real industrial construction projects from a partner contractor.
No takes yet. Share an insight, caveat, or question.
Ali et al. (2016) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: