PCL Tutorial at ICRA 2013

When: May 10, 2013

Where: International Conference on Robotics and Automation (ICRA) 2013, Karlsruhe, Germany.


Short Description:

The number of research initiatives that employ 3D point cloud processing has increased substantially since the introduction of affordable off-the-shelf 3D cameras, such as the Microsoft Kinect. This is shown, among other things, by the number of workshops on related topics in the fields of robotics and computer vision. The proposed tutorial will equip researchers, developers and R&D engineers with the tools necessary to start realizing new and interesting ideas, based on the emerging Point Cloud Library (PCL), which builds on its large set of tools and state-of-the-art algorithms to allow advanced 3D processing. The goal is to provide an overview of standard solutions to common 3D processing problems and especially 3D object recognition. In particular, we will guide our guests through the development of a complete object perception pipeline during the event, employing modules such as filtering, feature estimation, segmentation, registration, and object recognition. As during the previous tutorials, we will encourage the audience to bring their own Microsoft Kinect or PrimeSense sensors, so that they can try out the presented steps directly and run the demos on their own laptops. The invited speakers will talk about the features in PCL, present (using live demos) how they used it their work, and will be available to answer spontaneous questions of the participants during the sessions and in separate conversations during the brakes.

List of speakers:


9:00 - 9:10 F. Tombari, D. Holz Introduction
Session #1: Basics
9:10 - 9:35 J. Sprickerhof “We want your code, or how to develop for PCL”
9:35 - 10:00 D. Holz PCL Basics
10:00 - 10:30 Coffee Break
Session #2: Object Modelling
10:30 - 10:55 F. Tombari "How does a good feature look like?"
10:55 - 11:20 S. Miller "Clean Models from Noisy Sensors: Registration and Reconstruction Techniques"
11:20 - 11:50 D. Holz, J. Sprickerhof "3D Registration with PCL"
11:50 - 14:15 Lunch Break
Session #3: Segmentation and Recognition
14:15 - 14:40 A. Trevor "Organized Point Cloud Segmentation"
14:40 - 15:05 A. Aldoma "3D Object Recognition and 6DOF Pose Estimation"
15:05 - 15:30 Z.-C. Marton "Ensemble Learning for Object Recognition and Pose Estimation"
15:30 - 16:00 Coffee Break
Session #4: People detection
16:00 - 16:25 K. Buys "People detection on GPU"
16:25 - 16:50 M. Munaro "People Detection in RGB-D Data"