| Paper ID | C-3-2.5 |
| Paper Title |
3D Point Cloud Labeling Tool for Driving Automatically |
| Authors |
MingHui Li, Shenzhen Unity-Drive Innovation Technology Co, Ltd, China; Yanshan Zhang, Zhengzhou University of Aeronautics, China |
| Session |
C-3-2: Machine Learning and Data Analysis 2 |
| Time | Thursday, 10 December, 15:30 - 17:15 |
| Presentation Time: | Thursday, 10 December, 16:30 - 16:45 Check your Time Zone |
|
All times are in New Zealand Time (UTC +13) |
| Topic |
Machine Learning and Data Analytics (MLDA): |
| Abstract |
LIDAR（light detection and ranging） and visual perception are the key factors for the success of high level (L4-L5) automatic pilot obstacle avoidance.The combination of deep learning and 3D point cloud undoubtedly lays a solid foundation for the rapid development of automatic driving.At the same time, the demand of a large amount of data urges us to improve and perfect the point cloud marking tools.This article describes a newly developed 3D point cloud annotation tool, it supports PCD and bin formats. Using point cloud tracking P2B algorithm to achieve semi-automatic labeling, and using the reference of the z-axis heading Angle automatic detection function to simplify the complexity of the pull frame, It achieves the conversion of 3D bounding box coordinate information to 2D bounding box coordinates of point clouds and images acquired after the joint calibration of camera and lidar. It simplifies the operation of labeling tools and improves the efficiency of labeling. |