Dr Dorian Tsai

Associate Investigator and Post-doctoral Research Fellow

Dorian is passionate about working with scientific researchers and industry to solve key enabling problems in environmental robotics with application to large-scale management and monitoring challenges. Dorian is currently an Associate Investigator and Research Fellow for the Reef Restoration and Adaptation Project (RRAP) at the QUT Centre for Robotics. His research expertise is in robotic vision, operating at the intersection between deep learning, computer vision and robotics.

In RRAP, he is the Tech Lead in the Translation to Deployment sub-program’s Technology Development stream. He is looking at research and technological interventions aimed at helping to save the Great Barrier Reef from climate change. In particular, his focus is on identifying, developing and engineering robotic technologies to improve coral monitoring processes in large-scale aquaculture facilities. Additionally, Dorian has developed machine-learning pipelines to automatically track endangered sea turtles from drone imagery around Raine Island. Previously in an agricultural robotics project, Dorian collaborated with academic and industry partners to automatically detect various weed species for precision pasture and vineyard management using state-of-the-art machine vision and deep learning techniques.

Dorian received his Bachelor’s of Applied Science in Engineering Science from the University of Toronto, Canada, and a double Masters of Science in Space Science & Technology from Lulea Technical University in Sweden, and in Robotics & Automation from Aalto University in Finland. He completed his Master’s thesis project at the NASA Jet Propulsion Lab on autonomous vision-based tether-assisted rover docking. He then worked as a robotics engineer at the Canadian Space Agency before arriving at QUT for his PhD. Dorian completed his PhD at QUT with Peter Corke, Donald Dansereau and Thierry Peynot. Dorian investigated 4D light field features and how to exploit them for camera motion and vision-based control for robots, enabling improved structure-from-motion around transparent objects, where traditional methods would otherwise fail.

Projects