Over the summer of 2018 I interned abroad in Shanghai at SAIC's Vehicle Software & Intelligence Center.

During my internship I developed a collection of ROS (Robot Operating System) nodes that utilized PCL (Point Cloud Library) to interpret point cloud data. The program was capable of reporting location and type of various objects above the road surface. Below a before and after image can be seen of how my program visualized the objects it detected on the road using RVIZ. Each vehicle has a box around it originating at its estimated center. The color of the box depends on an estimate of the type/size based on number of points and proximity. The green rings represent the points the program things are on the road plane. To detect the objects the program utilizes clustering based on difference of normals for points above the road plane. Using well designed launch files and ROS parameters the program can quickly be set to work with data from various 3D lidars, .PCD files, and ROS bag streams.

During my internship I learned a lot about lidars, PCL, ROS, C++, and Chinese culture.

Project Github Repository:

https://github.com/niko1499/Lidar_Utility

Company Article about internship:

https://mp.weixin.qq.com/s/ZfcZE6RBF-r7RCSsfABUzQ