Research On The Construction Of Multi-lod BIM Models For Typical Components Of Song-style Dou-Gong Based On Point Cloud

Wooden architectural heritage(WAH)inherits brilliant culture,which is one of the most important immovable cultural relics in China.For the assessment and improvement of the safety performance as well as historical and cultural inheritance of WAH,the dimension information of various typical components of such architecture is the important foundation.Aimed at the key and difficult problems in the protection and development of WAH,which is efficient and high-precision acquisition of characteristic parameters of multiple-level-of-detail(multi-LoD)parametric mathematical model of typical components,taking the typical Chinese WAH constructed in the Song Dynasties as an example,the paper focuses on the standardized component features of typical components,multi-LoD parametric mathematical models for different requirements.and a method for automatically extracting the multi-LoD parameters.The main research contents and conclusions are as follows:(1)Based on the construction method,the standardization characteristics of the typical components of ancient wood buildings are summarized.The modular construction methods and shape features of the components are summarized,which lays the foundation for the multi-LoD parametric mathematical model(including size and shape)for different requirements.(2)Aimed at the different requirements in different stages of WAH,combined with the standardization characteristics of the typical components,a framework of multi-LoD parametric models,including corresponding parameters and shape grammars,is developed and implemented in BIM.This provides a clear target for efficient and accurate feature extraction based on high-fidelity point cloud data.(3)A method for automatically extracting the abovementioned parameters is developed and the critical parameters of this method are recommended.The method consists of a point-cloud slicing algorithm,outline-extraction algorithm,characteristic-point detection algorithm and a characteristic parameter extraction algorithm.(4)The typical components are verified by examples,and the results indicated that the relative and absolute errors of such parameters were less than 5% and 1 mm,respectively.Among them,the absolute error of 93% is less than 0.5mm,and the relative error of 89% is less than 2%.In addition,this method was capable of extracting parameters from millions of point-cloud data within 7 minutes;thus,validating the high efficiency and reliability of the proposed method.Furthermore,the LoD300 BIM model was established using the corresponding parameters,and the corresponding maximum deviation is less than 1.8mm,average deviation is less than 0.3mm,standard deviation and RMS estimate are less than 0.5.The deviation is small overall,even negligible;thus,validating the rationality and reliability of the proposed parametric model.