Livox Simu-dataset contains point cloud data and corresponding annotations generated based on the autonomous driving simulator, and supports 3D object detection and point cloud semantic segmentation tasks. The sensor contains 5 Horizon lidars and 1 Tele-15 lidar. The entire dataset contains 14,445 frames of 360° Lidar point cloud data, 3D bounding box annotations for 6 types of objects, and semantic annotations for 14 categories. The simulated scene contains a variety of vehicle and pedestrian models and corresponding traffic flow. In addition, a variety of traffic lights, traffic signs, partitions (isolated railings, green belts, isolated piers, etc.), trees and buildings make the entire scene closer to the actual driving scene.
Data format and download link:
PLEASE READ THESE TERMS CAREFULLY BEFORE DOWNLOADING THE DATASET. DOWNLOADING OR USING THE DATASET MEANS YOU ACCEPT THESE TERMS: GNU General Public License v3.0
Object detection demo: