PDAL — The Complete Guide: Processing Pipelines for LiDAR and Scan Point Clouds
PDAL is an open-source library and command-line toolset (BSD-3-Clause licence) for translating and processing point clouds. Think of it as the GDAL of LiDAR and scan data: there is no graphical interface, but it is a powerful tool for automated processing.
PDAL
PDAL contributors (Hobu)
Point cloud processing pipelines for LiDAR — filter, classify, crop and convert LAS/LAZ, like GDAL for point clouds.
Get it from the vendorWhat it is used for
- Reading and converting — between point cloud formats, including LAS and LAZ.
- Noise removal — removing outlier points.
- Ground classification — separating the terrain from vegetation and buildings.
- Cropping by polygon — just the project area.
- Thinning — reducing heavy files to a workable size.
How to use it
You build a pipeline of stages — read, filter, classify, crop and write — and run it on one file or on hundreds. That is how you produce terrain models from scans and process drone surveys. Work is done on the command line or in Python, so it helps to know one of them.
Related tools
For viewing and editing point clouds in a graphical interface — CloudCompare. For sharing a point cloud in the browser — Potree. For producing a point cloud from drone photos — WebODM.
CloudCompare
CloudCompare Project
Point cloud and laser scan processing — registration, cleaning, as-built comparison and measurement.
Get it from the vendor
GDAL
OSGeo
Command-line tools and a library to convert, reproject, clip and process geodata — DEMs, GeoTIFF, Shapefile and more.
Get it from the vendor
Potree
Markus Schütz / TU Wien
Browser viewing of huge LiDAR and scan point clouds, with distance, area, volume and height-profile measurements.
Get it from the vendor
WebODM (OpenDroneMap)
OpenDroneMap
Photogrammetry from drone imagery — orthophoto, elevation model, point cloud and textured 3D model.
Get it from the vendorThe bottom line
PDAL turns point cloud processing into a repeatable process. For anyone who receives scans regularly, a pipeline written once saves a great deal of manual work.