LiDAR Processing: 5 Steps from Raw Point Cloud to CAD Deliverable

LiDAR processing is the work between the scanner and the deliverable: importing raw LiDAR point clouds, checking them, cleaning and classifying them, extracting what the project needs and handing it to CAD. Whether the data comes from an aerial survey, a drone, a terrestrial scanner or a mobile mapping system, the steps are the same.

1. Import and inspect. A good LiDAR software opens the point cloud as it is, even when a project reaches several billion points. VisionLidar uses tiling so that very large LiDAR point clouds stay smooth to navigate on a standard Windows workstation.

2. Clean and classify. Classification separates ground, buildings, vegetation and noise so each class can be used on its own. VisionLidar automates ground, building, vegetation and noise classification and can segment data into up to 128 classes, then run object detection in batch.

3. Extract features. Surveyors and engineers rarely deliver points: they deliver objects. In VisionLidar you measure, vectorize and extract poles, curbs, road markings and power lines directly in the point cloud, and store them as points or polylines.

4. Produce deliverables. From the classified ground you generate TIN surfaces, contour lines, profiles, sections, alignments and volume calculations in the same point cloud software, without switching tools.

5. Hand off to CAD and share. The extracted features go to AutoCAD, Civil3D, BricsCAD or MicroStation. For teams and clients who do not have a LiDAR software installed, projects can be published and shared in a browser with VL365.

To see these LiDAR processing steps on your own data, request a VisionLidar demo or download the free VisionLidar Viewer from the VisionLidar page.