Point Cloud Toolbox
R2026bPoint Cloud Toolbox™ provides algorithms and apps for designing and analyzing point cloud processing systems. The toolbox provides functions for registration, filtering, geometric transformation, clustering, and segmentation of point clouds. It supports processing point cloud data from sensors such as lidar, RGB-D cameras, stereo cameras, and mmWave radar, and from workflows such as visual SLAM, photogrammetry, and structure-from-motion.
You can read data from standard formats or stream it directly from lidar sensors. Visualization tools support analysis, measurement, and interactive editing of large point clouds.
Apps enable lidar-camera calibration, multi-sensor labeling, and registration. You can apply deep learning techniques for the detection, segmentation, and classification of point clouds. You can also generate C/C++ (with MATLAB® Coder™) and CUDA® codes (with GPU Coder™) for deployment.
Get Started
Learn the basics of Point Cloud Toolbox
Import, Export, and Visualization
Read, write, and visualize lidar point cloud data, process large point clouds
Filtering, Conversion, and Geometric Operations
Process point clouds with filtering, conversion, meshing, transformation, and geometric model fitting
Labeling, Segmentation, and Detection
Label, segment, detect, and classify objects in point cloud data using deep learning and geometric algorithms
Registration and SLAM
Register point clouds using algorithms, such as ICP or NDT, or feature-based techniques, implement SLAM algorithms with 3-D point cloud data or 2-D lidar scans
Calibration and Sensor Fusion
Perform lidar-camera calibration by finding extrinsic parameters between sensors, fuse data between sensors
Lidar Data Acquisition and Sensor Simulation
Acquire lidar data from supported third-party hardware, create synthetic lidar sensor measurements for simulation
