OpenPCDet
lastmod 2026-11-10created 2026-04-172 edits
OpenPCDet is a PyTorch-based open-source toolbox that collects models for detecting 3D objects in LiDAR point clouds. Representative models such as PointPillars, SECOND, PV-RCNN, and PointRCNN can be trained and evaluated within the same structure, and it provides loaders for public datasets such as KITTI, nuScenes, and Waymo. Data and model configuration are separated into YAML settings, making it easy to experiment with new models, and together with MMDetection3D it is a widely used codebase in this field.
OpenPCDet Toolbox for LiDAR-based 3D Object Detection.