摘要
建筑信息模型(BIM)在城市更新与既有建筑改造中的作用日益凸显,基于点云的BIM建模方法,因其高精度、高时效性的优势,而广泛应用于建成环境的建模。然而,既有建筑室内空间布局复杂、构件细节繁多,依赖点云的逆向建模存在原始数据冗杂且缺乏语义信息,导致建模过程耗时费力,难以准确还原实际场景。为此,基于点云语义解析,提出既有建筑室内BIM自动化逆建模方法。利用两阶段配准与组合降噪算法对原始点云数据进行精细化处理;通过改进的点云实例分割算法实现墙、板、门、窗等室内构件的细粒度分类,并结合构件几何形状拟合与参数提取完成空间语义解析;设计参数化逆建模方法,解决点云与BIM之间的关联问题;提出四维度模型质量评价方法,定性与定量地评估重建效果。以某大型酒店室内翻新工程为例,验证该方法的有效性。结果表明:所提方法可以实现95%的重建对象房间数量,重建构件数量比例为84.9%,重建平均位置尺寸偏差15 cm,重建效率3.84 m^(2)/min,每100 m^(2)重建误差数量6.38个,验证了所提方法的有效性。
Building information modeling(BIM)is playing an increasingly important role in urban renewal and existing building renovations.Point cloud-based BIM modeling,due to its high precision and efficiency,is widely applied for built-environment modeling.However,the complex spatial layouts and detailed components of interior spaces in existing buildings pose major challenges for reverse modeling,including redundant data and insufficient semantic information,making the process labor-intensive and less accurate.This study proposes an automated BIM reverse modeling method for existing building interiors based on point cloud semantic analysis.A two-stage registration and combined denoising algorithm is used to refine raw point clouds,and an improved instance segmentation approach enables fine-grained classification of walls,slabs,doors,and windows.Geometric fitting and parameter extraction support spatial semantic interpretation.A parameterized reverse modeling pipeline is developed to bridge point cloud data with BIM.A four-dimensional evaluation framework is introduced to assess reconstruction quality both qualitatively and quantitatively.The approach was validated in an interior renovation project of a largescale hotel.Results showed a reconstruction coverage of 95% for room-level objects,84.9% for building components,an average positional deviation of 15 cm,modeling efficiency of 3.84 m^(2)/min,and reconstruction error rate of 6.38 items per 100 m^(2),demonstrating the method’s effectiveness.
作者
卢昱杰
王硕
王明康
赵宪忠
LU Yujie;WANG Shuo;WANG Mingkang;ZHAO Xianzhong(College of Civil Engineering,Tongji University,Shanghai 200092,China;Key Laboratory of Performance Evolution and Control for Engineering Structures of the Ministry of Education,Tongji University,Shanghai 200092,China;Shanghai Research Institute of Intelligent Science and Technology,Tongji University,Shanghai 200092,China)
出处
《建筑结构学报》
北大核心
2025年第9期119-131,共13页
Journal of Building Structures
基金
国家重点研发计划(2022YFC3801700)
上海市科技创新行动计划(22dz1207100,22dz1207800)
中央高校基本科研业务费专项资金资助(2024-1-ZD-02)。
关键词
城市更新
建筑信息模型
点云语义解析
建筑室内构件
自动化逆建模
urban renewal
building information modeling(BIM)
point cloud semantic analysis
components
automated reverse modeling indoor building
