Sensing
Robots that see in all conditions
Research
Overview
Robust Vision
The Robust Vision program develops robotic vision algorithms and novel vision hardware to enable robots to see and act in all viewing conditions. The program is developing a suite of algorithms that enable robots to perceive their environments and consequently act purposefully under the incredible range of environmental conditions possible, including low light, rain, snow, ice sleet, fog, smoke, dust, wind, glare and heat. In addition, it is further developing innovative sensing hardware to facilitate robot operation under challenging viewing conditions such as low light, or through partial obscuration cameras and hyperspectral cameras.
The key question we are addressing is, how can innovations in existing computer vision and robotic vision techniques and vision sensing hardware enable robots to perform well under the wide range of challenging conditions encountered by robots and applied computer vision technology in the real world?
People
-
Michael Milford
Chief Investigator, Robust Vision Program Leader, RV 1 Project Leader
-
Queensland University of Technology
-
Chuong Nguyen
Research Affiliate
-
CSIRO DATA61, Canberra
-
Hongdong Li
Chief Investigator, Algorithms & Architecture Program Leader
-
Australian National University
-
Chunhua Shen
Chief Investigator, VL2 Project Leader
-
University of Adelaide
-
Jonathan Roberts
Chief Investigator
-
Queensland University of Technology
-
Peter Corke
Centre Director, Chief Investigator, QUT Node Leader, AA3 Project Leader
-
Queensland University of Technology
-
Niko Sünderhauf
Chief Investigator
-
Queensland University of Technology
-
Jürgen “Juxi” Leitner
Research Fellow, VA1 Project Leader
-
Queensland University of Technology
-
Donald Dansereau
Research Affiliate
-
Stanford University
-
Sean McMahon
PhD researcher
-
Queensland University of Technology
-
Sourav Garg
PhD researcher
-
Queensland University of Technology
-
James Mount
PhD researcher
-
Queensland University of Technology
-
Dorian Tsai
PhD researcher
-
Queensland University of Technology
-
Dan Richards
PhD researcher (currently on leave)
-
Queensland University of Technology
-
Timo Stoffregen
PhD researcher
-
Monash University
-
Medhani Menikdiwela
PhD researcher
-
Australian National University
-
James Sergeant
PhD researcher (currently on leave)
-
Queensland University of Technology
Projects
RV1: Robust Robotic Visual Recognition
Ongoing
Michael Milford, Hongdong Li, Jonathan Roberts, Chunhua Shen, Jürgen “Juxi” Leitner, Chuong Nguyen, Niko Sünderhauf, Sourav Garg, Sean McMahon, James Mount, James Sergeant, Medhani Menikdiwela
Robust robotic visual recognition for adverse conditions will develop algorithms that solve the fundamental robotic tasks of place recognition and object recognition under challenging environmental conditions including darkness, weather, adverse atmospheric conditions and seasonal change, and translate them into applications in industry. This project is relatively mature and hence is now pushing more heavily towards industry outcomes and engagement than some of the other projects. The key question we are addressing is, how can existing computer vision and robotic vision techniques be innovated to enable them to perform well under the wide range of challenging conditions encountered by robots and applied computer vision technology in the real world?
Previous project: RV2: Novel visual sensing & hybrid hardware for robotic operation in adverse conditions & for difficult objects
August 2016 - 2017
Chuong Nguyen, Peter Corke, Michael Milford, Donald Dansereau, Dan Richards, Dorian Tsai, James Mount, Timo Stoffregen
Novel visual sensing for robotic operation in adverse conditions will advance the performance of robot vision algorithms by using new conventional low light cameras, as well as rotational filters, hyperspectral cameras and thermal cameras to improve robot autonomy under any viewing condition. We are looking at the development of new algorithms that exploit the particular advantages of these innovative cameras to enable new performance benchmarks in tasks such as scene understanding, place recognition and object recognition. The development of hardware-software solutions that can deal with these corner cases - reflections, transparency and low light conditions - will be applicable to all Centre projects involving visual sensing and can be used by those projects to robustify their systems for performance under adverse conditions. The key question we are addressing is, how can existing and new special vision hardware be exploited and developed in conjunction with new algorithm development to enable robotic vision systems to operate well under the wide range of challenging conditions encountered by robots and applied computer vision technology in the real world?
Australian Centre for Robotic Vision
2 George Street Brisbane, 4001
+61 7 3138 7549