Every project starts with a broad idea. Initially, I planned to develop an autonomous robot for construction site inspections, using LiDAR and a camera to scan and verify structures. However, after discussing it with my supervisor, we agreed that the scope was too ambitious given the available time. Instead, we refined the project to something more focused and achievable—a Lidar-based AI system for construction site verification.

Research and Key Techniques

Before narrowing things down, I looked into different methods used in construction verification:

• Manual Inspections – Still standard but slow, labour-intensive, and prone to errors.

• Photogrammetry, which utilises Drones and cameras, is suitable for large sites but less effective indoors due to lighting and occlusions.

• LiDAR Scanning – Creates high-accuracy 3D maps that can be compared with blueprints and used by companies like Doxel.

• AI for Structural Comparison – Automatically identifies differences between scans and floor plans, reducing human effort.

Final Project Focus

After a second meeting with my supervisor, we dropped the autonomous robot aspect and focused on the Lidar-based verification system.

The core goals are:

• Capture and process 3D Lidar data from a construction site.

• Compare scans to DXF floor plans to verify the presence of missing or incorrect structures.

• Export processed data in practical formats for further analysis and review.

This keeps the project realistic while still addressing a real industry challenge. Next, I’ll start setting up the Lidar system and testing data processing. In my next blog, I’ll explain why we moved away from the robot idea and how I’ll approach the comparison system in more detail.