Point Cloud Library (PCL)

An open-source C++ library for processing 3D point clouds — filtering, registration, and segmentation included.

PCL is an open-source C++ library for point-cloud processing, originally started at Willow Garage and formally introduced in a 2011 ICRA paper, released under the BSD license. It packages the common steps of point-cloud processing into modules: voxel downsampling and outlier removal (filtering); computing features like surface normals and FPFH; ICP and NDT registration; RANSAC plane fitting and clustering-based segmentation; and surface reconstruction. Depth cameras and lidar both output point clouds, so PCL has long been a foundational tool for 3D perception in robotics, and ROS interconverts with it through the perception_pcl package. If most of the work is happening in Python, Open3D is a lighter-weight alternative.

ExampleUse PCL's RANSAC to fit and remove the tabletop plane, then run Euclidean clustering on the remaining points to get a separate point cloud for each object on the table.

Also called

PCL

Related

Point Cloud · Open3D · Iterative Closest Point · Point Cloud Registration · Point Cloud Segmentation · OpenCV (Open Source Computer Vision Library)

Sources

Point Cloud Library (PCL)
PointCloudLibrary/pcl - GitHub