BIM_Updating

This repository contains a complete implementation method proposed in the paper for semi-automated local updating for as-built BIM of piping systems using point cloud data.

Semi-Automated As-Built BIM Updating for Piping Systems

This repository contains the complete implementation of methods proposed in our paper for semi-automated local updating of as-built BIM of piping systems using point cloud data.

🎯 Overview

Our approach addresses the challenge of maintaining accurate as-built BIM models by automatically detecting and quantifying geometric changes in piping systems through point cloud analysis.

Key Features

• Automated piping segmentation: Direct extraction of piping networks from point clouds
• Geometric change detection: Quantification of changes in length, height, radius, and angle
• As-designed vs As-built comparison: Systematic analysis of design deviations

🏗️ Methodology

1. Point Cloud Preprocessing

• MATLAB Version:
  • Data processing for as-built point cloud models
  • Planar object segmentation
  • Multi-elevation piping system segmentation
• PointNet++ Version:
  • Custom-trained network for direct piping segmentation
  • Real-time piping network extraction

2. Spatial and Topological Analysis

• Geometric Parameter Extraction: Automated measurement of pipe dimensions
• Change Quantification: Statistical analysis of geometric deviations

3. As-built BIM generation

• As-designed Processing: Revit model parameter extraction by using Dynamo
• As-built Processing: Point cloud-based geometric analysis

🚀 Quick Start

Prerequisites

# Python environment
conda create -n bim-updating python=3.8
conda activate bim-updating

# MATLAB (for spatial analysis module)
MATLAB R2023b or later with Computer Vision Toolbox

📊 Results

Our method achieves:

• Segmentation Accuracy: 96% for piping system segmantation
• Geometric Precision: ±5mm for dimensional measurements
• Processing Speed: 70% faster than manual methods

🔧 Technical Details

Algorithms

• PointNet++: Modified architecture for piping-specific features
• RANSAC/ICP: Robust geometric fitting and alignment
• Connected Components: Instance segmentation of individual pipes

Data Formats

• Input: PLY, PCD, LAS point clouds
• Output: JSON geometric parameters, CSV change reports
• Visualization: 3D interactive models

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

🙏 Acknowledgments

Real Point Cloud Data: Self-collected by the authors using laser scanning equipment Synthetic Point Cloud Data: Generated using Blensor simulation framework Real BIM Models: Self-constructed Revit models by the authors Simulation BIM Models: Sourced from SimAUD Dataset