Research on Lightweighting and Rendering Optimization Techniques for Building Information Models

Abstract

Traditional Building Information Modeling (BIM) services, using a client-server (C/S) architecture, struggle to adapt to the new demands and changes in the field of architectural informatization. The development of a comprehensive BIM visualization system, which is web-based, requires no installation or downloading, is cross-platform, and facilitates easy sharing, has become a new pathway for BIM advancement. However, when rendering the vast amount of data in BIM three-dimensional scenes on the web, several issues persist, including low transmission efficiency, slow loading speed, laggy screen display, and low rendering frame rates, particularly on mobile devices and terminals with limited hardware performance. Therefore, this paper proposes BIM lightweighting techniques based on geometric data and texture information, as well as a web-based rendering optimization method for three-dimensional models. In terms of BIM lightweighting, more efficient methods and algorithms are employed to simultaneously lighten the geometric data and texture information of the models. For rendering optimization, an efficient view frustum culling algorithm is introduced, along with a design for an adaptive rendering strategy to enhance rendering performance.Through testing the loading efficiency of different scale BIM models before and after optimization was tested on the web. Results show that, while maintaining display accuracy, the average loading time of the optimized models was reduced by 40% compared to the unoptimized models. The average data compression rate reached 45%, and the average memory usage decreased by 32.55%. After stable rendering, the frame rate was close to 60fps.