CSSE Seminar Series

"Improving 6 DoF Pose Estimation for Agricultural Environments"

Speaker

Guillaume Payet

Institute

University of Canterbury

Time & Place

Mon, 19 Oct 2020 15:00:00 NZDT in E16 - Engineering Core

Abstract

6 degree-of-freedom pose estimation allows augmented reality applications and autonomous machines to know the location and orientation of objects of interest. Current research in 6D pose estimation focus on indoor robot manipulation. Our research will investigate the performance of existing methods on outdoor scenes with complex, natural objects and human-like motion. For this research, we will develop a new dataset consisting of grape vines in an orchard filmed from a head-mounted camera. After evaluating the existing methods, our own novel method will also be tested and evaluated. We expect our method to perform better than the current state-of-the-art while requiring less information from the human-labelled datasets.

Biography

Guillaume Payet is a PhD student under the supervision of Prof. Richard Green. He graduated with a BSC(Hons) in 2016 from the University of Canterbury. He went abroad to work and decided to come back to do a PhD. Guillaume's interests are in computer vision, deep learning, augmented reality, virtual reality and game development.