Hybrid Procedural Modeling System of Lingnan Ancestral Hall

Procedural modeling has become increasingly applied in the architecture field to address the high costs of manual methods in digitalizing architectural heritage. However, traditional procedural modeling faces challenges with complex and difficultto-adjust parameters. Therefore, this study introduces a hybrid procedural modeling system using Lingnan ancestral halls as a case study. Lingnan ancestral halls, with their intricate architectural styles, present challenges for traditional modeling methods. Our system combines digital image processing with procedural modeling, utilizing an improved segmentation algorithm based on down-sampling to process floor plans and record room dimensions and locations. A k-means clustering algorithm classifies identified rooms to determine key features such as roads and entries, guiding courtyard layout generation. Based on these layouts, the system procedurally generates the ancestral hall model, allowing users to fine-tune parameters for accuracy. Experimental results demonstrate that this system effectively addresses the complexity and inflexibility of parameter adjustments in traditional methods, making it a valuable tool for the digital preservation and reproduction of Lingnan ancestral halls.