Research on Building Space Model Method Based on Big Data Map Visual Design

Abstract

In order to explore the architectural space model of design, a visualization method based on big data map is proposed. Referring to the tile pyramid model, a multidimensional aggregation pyramid model (MAP) is proposed, which extends the 2D spatial hierarchical aggregation of tile pyramid to the multidimensional of time/space/attribute and supports the multidimensional hierarchical aggregation of time, space, and attribute. Then, taking spark cluster as the parallel preprocessing tool and HBase distributed database as the persistent storage of map model data, an open-source distributed visualization framework (MAP-Vis) is realized. Then, the BIM model is reconstructed, and a component instance hierarchical splitting strategy based on IFC structure tree is proposed to separate the digital and analog of the original IFC file. The reconstructed IFC model file is transformed into glTF format file, and the dual relationship mapping of geometric space and semantic attributes is completed in the transformation process. Finally, the visibility detection algorithm of BS-AB scene components based on the hierarchical bounding volume (BVH) structure is proposed to eliminate the visibility of building components. The experimental results show that BIMviews is slow to load the IFC file of the experimental object and obtain the model data, with an average of about 40 s, and the Caton is obvious. However, it only takes about 7 s to load glTF file into big data map visualization design by Three.js. It is verified again that glTF format is more suitable for BIM model data than IFC format. The visualization design, display, and interaction based on big data map are based on glTF format. It proves the effectiveness of big data map visualization.