Robotic Vision Quarterly Newsletter - February 2017
View this email in your browser

creating robots that see


Centre researchers competed in the Maritime RobotX Challenge in Hawaii in December and came in 2nd place



Welcome to our second public newsletter and the first for 2017. We hope that this is an informative and enjoyable way to keep up to date with what’s happening in our Centre's quest to create robots that see.

Since the last newsletter we have been working hard to better articulate our purpose.  This is important not just for explaining what we do, but also for shaping our research program and as a lens for our researchers to see the big picture that their research fits into.  We have set ourselves three grand challenges: environment, healthcare, and resources.  Specific examples of opportunities associated with these grand challenges include inspection of our natural and built environments, medical robotics and sustainable food production.  The Centre’s updated vision and mission, including grand challenges, can be found in this infographic.  Articles in the newsletter below introduce, or provide updates on, projects that contribute to solving our grand challenges.

A complete snapshot of last year's progress is being compiled into our 2016 annual report and this will be available in April.  Previous annual reports can be found on the Centre's website.  More timely information about Centre achievements can be found on our website and our YouTube channel. We are undertaking a redesign of our website to better portray our activities, and this should be up and running by the time of the next newsletter.

We believe that the ability to see, to visually understand the complex world around us and respond to it, is critical for the next generation of robots that will perform useful work in agriculture, environmental monitoring, healthcare, infrastructure inspection, construction, manufacturing and so on. If you have problems that robotic vision might solve, please get in touch. If you’d like to help us translate our science into impact and wealth creation, please let me know.
If you have comments about the newsletter, questions about what we do or how to engage with us, please feel free to email me.
Enjoy the issue.

Professor Peter Corke
Centre Director



With the support of PwC we're running a series of innovation challenges aimed to identify people and companies operating in a similar space to us in the Australian robotics and vision communities.

We have two aims:

(1) to identify companies we can collaborate with on solving challenges, partners who we can work with or already have solutions to some of the problems we are asked to address, and companies who might be interested in working with us to commercialise technologies in the future.

(2) to engage with companies who broadly operate in the automation/robotics/sensing/vision space to collectively develop a robotic and vision roadmap for Australia and raise awareness of the important economic benefits of the industry to Australia.

We see the innovation challenges as an opportunity to partner on robotic vision solutions for different sectors and to raise the profile of the robotic vision industry. ​

Our first two challenges are focused on robotic vision solution providers in the manufacturing and agriculture sectors.

An information session around the manufacturing challenge will be held in Brisbane on 10 March 2017, while the food production challenge will be held in Brisbane on Wednesday 21 June 2017.

The Centre is particularly keen to seek involvement from all companies in Australia broadly operating in the robotics, computer vision, automation and sensing space as well as manufacturers of robotics, vision and sensing equipment, tech companies and industry bodies.

By running these challenges we hope to:
  • Create a network of companies in the robotic vision space in Australia
  • Identify potential collaboration opportunities
  • Raise industry profile with State and Federal governments, with a view to gain increased focus on the industry
  • Gain recognition of the importance of robotics and vision to the Australian economy
  • Increase competitiveness with similar industries overseas
For further information please
contact the Centre
LIEF Grant provides funding for new Deep Learning Supercomputer

Deep learning is a term that has not been far from anyone's lips since the inception of the Centre, with its impact on computer vision and robotics in the last 2 years both immediate and profound.  

A new deep learning supercomputer that just went online at The University of Adelaide is now providing the computing power the Centre needs to train deeper and larger models and to handle large numbers of parameters and big data.

Three Centre nodes, Adelaide, QUT and ANU, along with Deakin and UWA were all a part of a successful LIEF (Linkage Infrastructure Equipment and Facilities) grant award by the Australian Research Council.  The grant, combined with funding from each of these universities, is worth nearly $400-thousand.

Centre Deputy Director Ian Reid was the driving force in getting the LIEF grant.

“The fact that we now have access to a computer that can train models, do evaluations, and so forth, is very significant for us,” Ian said, “because we would not be able to do much of what we’re trying to do, particularly in semantic vision, but in all sorts of other areas of the centre, were we not have access to the best models and the best recognition schemes.”

In addition to the deep learning supercomputer at Adelaide, the money is also being used to enhance computing power at QUT and ANU.


New Associate Investigator Associate Professor Felipe Gonzalez is leading pioneering research with QUT remote sensing researchers and Unmanned Aerial Vehicle (UAV) engineers. The team are using new miniaturised hyperspectral cameras to monitor the health of Australian landscapes. This new face of environmental monitoring combines UAVs and a highly-specialised camera that was previously so big and expensive that only satellites and airplanes could carry them. The team have just completed a data-gathering mission on the pristine Ningaloo Reef at the Ningaloo Coast World Heritage site in Western Australia.  The team were among the first in the world to obtain aerial hyperspectral imagery of a coral reef in extraordinary resolution. Media release here

Our ANU colleagues were successful in the 2016 Australian Renewable Energy Agency (ARENA) grant outcomes. The project titled “A Robotic Vision System for Automatic Inspection and Evaluation of Solar Plant Infrastructure” received a total of $3.1 million in funding.

The project is led by Associate Investigator Professor Fatih Porikli who is excited about this project, saying "using drones for automated inspection is a positive step towards a more efficient and productive society by leveraging on recent improvements in robotics and intelligent computer vision technologies.”

"This is the first time in the world that a flock of drones carrying optical sensors, such as optical, polarization and thermal cameras, laser scanners, empowered by intelligent computer vision solutions will be used for automated inspection of solar power plant facilities".

“It provides a perfect balance between the three pillars of innovation: fundamental scientific challenge, necessity arising from real problems, and opportunity to have a direct impact in wealth creation.”

The full media release is available 


Centre Associate Investigator Jason Ford


It’s a daunting task – having to maintain a network of hundreds of thousands of miles of power poles. That’s the challenge facing Queensland energy giant, Ergon.

Centre researchers are now part of an effort to help Ergon with its challenge. They are involved with a new CRCSI and QUT collaboration with Ergon to develop a vision and sensing platform for UAV’s to fly over power poles and capture images. They are looking for things like rot in a pole’s cross-arms, along with other possible structural issues.

Right now, the work is being done by UAV operators, who must keep the airborne vehicles within their line of sight. It’s hoped this project will make their job much easier, by using on-board sensors to develop a force field, if you will, that will keep the UAV’s from getting too close to the poles or wires.

Centre Associate Investigator Jason Ford is among those involved with the project, which is an example of the type of research being done to tackle issues with infrastructure.

“This project was empowered by the Centre existing and provides an example of the type of work the Centre can do,” Jason said.

Other Centre members involved with the project include Centre Director Peter Corke, Research Fellow Feras Dayoub, and Centre Research Engineer Steven Martin. Other researchers involved include Andrew Keir, from QUT’s Institute of Future Environments and QUT Research Fellow Aaron Mcfadyen.

Ergon is going through a merger with Energex, with the merged company to be called Energy Queensland.
QUT Postdoctoral Research Fellow Anjali Jaiprakash pictured at Imperial College London
Congratulations to Research Fellow Dr Anjali Jaiprakash

Anjali recently returned from Imperial College in London, where she was selected to take part in the Hamlyn Winter School on Surgical Imaging and Vision, December 2016.

In addition, Anjali and her team were able to scope an unmet clinical solution in the field of orthopaedics and present at the final day of the School in PechaKucha style.  The Hamlyn Centre also awarded Anjali the runner-up for the “Best Project Award” for her project on autonomous portal placement for knee arthroscopy.

“It was absolutely inspiring to learn and brainstorm solutions for unmet clinical needs from experts who are at the forefront of research in imaging, sensing and robotics for addressing global health challenges,” Anjali said. “In addition, my first time in London was heart-warming. I loved the arts and culture.”

The Winter School focused on both technical and clinical aspects of Surgical Imaging and Vision, with invited lectures, hands-on demonstrations, workshops, and mini-projects. 

Anjali received an Advance Queensland Research Fellowship in 2016. These awards are given to drive innovation and collaboration in new and existing industries and solidify the state’s capability and reputation as a global science and research leader.

Anjali’s work will make keyhole knee-surgery easy safe and predictable.

“As part of my Advance Queensland Grant,” says Anjali, “I am building a machine to move the leg safely to allow a robot to perform knee arthroscopy safely and reliably.”

The goal of her work is to eventually allow robots to perform the surgery autonomously. 

Anjali’s partners in the project already include one of the world’s leading medical technology firms, Stryker, along with Holy Spirit Northside Private Hospital and Prince Charles Hospital Foundation.

She is part of the Medical and Healthcare Robotics Group at QUT under the supervision of Centre Associate Investigators Robotics Engineer Professor Jonathan Roberts and orthopaedic surgeon, Professor Ross Crawford.

“This collaboration would not have been possible without the Queensland government recognising the importance of advanced technologies to the state economy and most importantly to the patients of Queensland,” said Anjali. “This grant will allow us to take this dream one step closer.”
Robo food: Harvey the harvester picking capsicums
Industry workshop
The Centre delivered an industry workshop in Toowoomba on 10 November 2016 to 35 people drawn from the agriculture industry.  The purpose of the workshop was to understand the issues and challenges that face their industry and with this understanding inform the research of the Centre. One of the key challenges that emerged from the workshop was inspection including issues such as identifying damage, pests and disease on crops.
Chris Lehnert was the star of the day not only giving an outstanding presentation to industry on agricultural robotics research but also did a stand up job in the media by doing a great piece to camera for Channel 7 Toowoomba.  You can see Chris’s interview
A special thanks goes out to Russel Rankin, Chair of the Centre’s End User Advisory Board who was instrumental in promoting the workshop to the industry.  Our thanks also goes to Richard Routley from the Department of Agriculture who kindly let us use their facilities and also to Stuart Hazell who took time out of his busy schedule to present programs available from AusIndustry.

Robo Food: Researchers at Centre Headquarters got to enjoy some robot harvested capsicums courtesy of Harvey and Chris Lehnert

Researchers during the knowledge leadership workshop held with the Centre's symposium RoboVis last year. Left to Right: PhD students Sean McMahon, Dorian Tsai, James Sergeant, Medhani Menikdiwela, and Research Fellow Chuong Nguyen

World first knowledge leadership program for early career researchers

We have an important responsibility to nurture the next generation of knowledge leaders, innovating robotic vision experts who will go on to work in industry, government and academia. Our research fellows and PhD researchers will help to translate robotic vision research to adapt existing industries and to convert robotic vision research into new products, services and enterprises so as to forge whole new industries.
To support our early career researchers we have commissioned a tailor-made knowledge leadership program from workplace psychologists,
Evexia. The program, which commenced last year, focusses on the skills our researchers need for successful careers. Unlike the plethora of other leadership training programs which focus on either the research or corporate worlds, our bespoke program covers some of the specialist skills required by researchers but also encompasses entrepreneurship, publicity engagement and resilience.  We hope to equip our researchers with skills that will help them to not only succeed, but to also lead in any sphere of influence.
The program started last year with the introduction of career development planning and has been supported by the Centre COO, Sue Keay, with workshops at each node on careers. The future focus of the program is on helping researchers with developing purpose and a unique identity, learning to influence, translating research and bringing entrepreneurial resolve to bear, and also on further career planning and resilience.
We have set the bar high and believe we have developed a unique program combining elements of traditional leadership training with a specific focus on research and entrepreneurship.

PhD student James Mount with his BBots

PhD student James Mount began shipping out his BBot robots in December. The BBots were created as part of a Kickstarter campaign, where James had to reach a goal of $500 to get the project up and going. He reached that goal in July 2016, a week before the Kickstarter deadline. He was then able to begin the design and assembly process.

James created 17 BBots to send to schools and individuals around Australia. BBots are easy-to-use, inexpensive robotic vehicles with several modes to help introduce kids to robotics and the scientific process. They don’t require any programming or mechanical design skills to get started.

James’ mother is a school teacher, so he wanted to figure out how he could give back and help schools.

“This allowed me to combine my family’s background in education with my expertise in robotics,” said James. “I think robotics is a great teaching tool that can engage and excite students.”

For four months at the end of 2016, James was his own one man assembly line, working in his spare time.

 “It's been a bit of a roller coaster. It’s been challenging at times, but in the end it’s been worthwhile. I hope that the people who supported the project are happy with the BBot, whether they want to use it to teach or just learn from.”

The goal of the BBot project is to engage kids and help them understand the scientific process. For example one of the worksheets uses BBot’s differential drive system to help kids explore the scientific concepts of hypothesising and experimental design. The workshop challenges students to think scientifically by asking them what they think BBot will do if both wheels are running at the same speed, if one wheel is running faster than the other, or if one wheel isn’t running at all.

James is working toward his PhD at the Queensland University of Technology (QUT). His PhD supervisor is Associate Professor Michael Milford, a Chief Investigator with the Centre..

Milford challenged James to take on the project, as a way to help schools deal with a world where technology is constantly evolving.

“I hope this is the start of an inclusive movement to give everyone the basic scientific and technological awareness they need to prosper in this rapidly changing world,” said Milford.

The project started with a very basic design, and then evolved to what it is today. James said he couldn’t be happier with how it turned out and his dream to provide an easily accessible robotics platform for schools has become a reality.
Centre Director Professor Peter Corke’s second FutureLearn short Massive Open Online Course (MOOC) has come to a close for a class of 10,600 students who learnt about simple robot kinematics over 3 weeks. The next course Introducing Robotics: Build a Robot Arm starts on the 20th February. More information is available on the FutureLearn website. A further 3 robotic vision MOOCs are being developed in 2017.
Our Robotic Vision Summer School is held at ANU's beautiful Kioloa Campus

Our third summer school will be held again at ANU’s Kioloa Campus from the 12th-17th March 2017. This international summer school targets Masters and PhD students, academics and industrial researchers. The week-long event is a great opportunity to learn about fundamental & advanced topics in robotic vision from our domestic and international researchers and includes talks, workshops, demonstrations and social events. Our 2017 speakers include Professor Davide Scaramuzza from the University of Zurich, Professor Javier Civera from the University of Zaragoza and Dr Simon Lucey, head of the CI2CV laboratory. Registrations close 17th February 2017 and more information can be found on the Summer School website
Congratulations to the whole QUT team including AI Matt Dunbabin, RF Markus Eich, and PhD students Peter Kujala and Will Chamberlain for the great result at the Maritime RobotX Challenge. The challenge is a biennial international autonomous surface vehicle competition and was held in Oahu, Hawaii from the 12th-18th December. The team placed 2nd in the 2016 competition winning $12,000.

Each team competes using standardised platforms (16ft WAM-V), custom fitted by competitors with a suite of sensors, computation, software, and hardware in order for the Autonomous Surface Vessel to perform a range of tasks in a course format, scoring points for demonstrating various competencies within a single “run” in 45-minute time window. Tasks included underwater survey (using a custom-built AUV), underwater acoustic pinger localization, navigation, obstacle avoidance, docking, and delivering projectiles within detected targets – all undertaken autonomously on an open-water course.

Teams are also required to author a journal-style paper describing the development, technology, and approach to addressing the Challenge, as well as a team video to introduce the team and their roles in developing the ASV. Furthermore, each team delivers a 20-minute presentation to the judging panel during the competition outlining specifics of their ASV and the team’s development toward meeting the challenges of the RobotX competition.
Some of the Centre researchers involved in the 2016 Amazon Picking Challenge
For a second straight year, the Centre is sending a team to take part in the newly-renamed Amazon Robotics Challenge. This year’s event will again be held in conjunction with Robocup in Nagoya, Japan at the end of July. The team is hoping to build off its success from last year’s Amazon Picking Challenge, where it finished sixth overall in the picking task out of the 16 teams from around the world.

“Last year, we essentially started from scratch and built a working system in just a few months. Now we’re looking at pushing this further,” says Centre Research Fellow Juxi Leitner.

Juxi led the Centre team last year, and is in charge of putting together this year’s team. For Juxi, this isn’t just about trying to solve the specific problems for this competition. He sees it as a chance to really advance the Centre’s mission of creating robots that can see and act robustly in real world scenarios.

“For us, robotic vision is not just computer vision on a robot. It’s the close interaction between vision and robotics,” Juxi says.

“The problems we face in this challenge allow us to bring vision and action together, creating something that is larger than its parts. Our hope then is to take what we learn from this and start applying it to other areas we’re trying to solve in the Centre, like agricultural, infrastructure and medical applications.”
Meet our Robots
AgBot II (with QUT research engineer Owen Bawden), Harvey (with Research Affiliate Chris Lehnert) and LunaRoo (with Research Fellow Juxi Leitner) featured in a recent Roboscope episode of Scope TV. Watch here

Paper recognised for Enduring Contributions to Robotics
Partner Investigator Frank Dellaert of Georgia Tech will receive the Association for the Advancement of Artificial Intelligence (AAAI) Classic Paper Award in February at AAAI 2017 in San Francisco, California. The award will recognize his work on the Monte Carlo Localization algorithm in an AAAI 1999 paper, Monte Carlo Localization: Efficient Position Estimation for Mobile Robots. In addition to Dellaert, co-authors on the paper included first author Dieter Fox, Wolfram Burgard, and Sebastian Thrun. Media release here

Gizmodo's Best Australian Science Discoveries of 2016
The Centre was featured in Gizmodo’s The Best Australian Science Discoveries of 2016 with the Rangerbot and Agbot II and Associate Investigator Professor Felipe Gonzalez and his work on tracking Koalas with AI-Equipped Drones.
Chief Investigator named AMP Tomorrow Maker
Chief Investigator Michael Milford was named as an AMP Tomorrow Maker 2016 and awarded $25,379 to help his Math Thrills program reach thousands of students Australia-wide. Media release here 
Centre Research Fellow Juxi Leitner being filmed for the Roboscope episode of Scope TV
Associate Investigator Felipe Gonzalez is leading research using Artificial Intelligence (AI) equipped drones to track koala populations
AMP Tomorrow Maker 2106 - Michael Milford
Chief Investigator Michael Milford talks about his Maths Thrills program designed to make Maths more thrilling for students every day

CI Hongdong Li from ANU

CI Robert Mahony from ANU

AI Tristan Perez from QUT
Recent Achievements

CI Hongdong Li has been appointed as an Associate Editor for the IEEE Transactions on Pattern Analysis and Machine Intelligence (IEEE-TPAMI).  This appointment was nominated by the IEEE PAMI's Advisory Board, and approved by PAMI Editor-in-Chief and Editorial Board. PAMI is the most highly regarded journal in the field of Computer Vision.  Measured by impact factor, PAMI is in fact the No. 1 journal over all publications in Computer Science and in Electrical Engineering.   PAMI is also the No. 1 most cited journal among all journals in EE and CS, and the no.1 in all IEEE Transactions.  CI’s Ian Reid and Hongdong Li are currently the only AEs from Australian Universities. More information about PAMI can be found here 

Chief Investigator Rob Mahony has been named IEEE Fellow in recognition of his contributions to control aspects of aerial robotics. Media release here

Chief Investigator Richard Hartley has been elevated to Fellow of AFCV (Asian Federation of Computer Vision) for his significant contributions to computer vision research especially in the areas of multiple view geometry and optimisation.

AI Tristan Perez has been appointed the Chair of the Technical Committee on Marine Systems of the International Federation of Automatic Control (2018-2021).

Best paper award DICTA 2016 to Zongyuan Ge, Chris McCool, Conrad Sanderson, Peng Wang, Lingqiao Liu, Ian Reid and Peter Corke for their paper ‘’Exploiting Temporal Information for DCNN-based Fine-Grained Object Classification’’

Best Student Paper Honourable Mention Award at ACCV 2016 to ANU student Tong Zhang and Associate Investigator Fatih Porikli for their paper "Sparse Coding on Cascaded Residuals"
Copyright © 2017  Australian Centre for Robotic Vision. All rights reserved.
As you are associated with the Australian Centre for Robotic Vision we have added you to our mailing list.

Our mailing address is:

Centre Headquarters
Australian Centre for Robotic Vision
S Block Level 11
QUT Gardens Point Campus
2 George Street
Brisbane QLD 4001

Add us to your address book

You can update your preferences or unsubscribe from this list


This email was sent to <<Email Address>>
why did I get this?    unsubscribe from this list    update subscription preferences
Australian Centre for Robotic Vision · S Block 1105 Gardens Point campus · 2 George Street · Brisbane, Qld 4001 · Australia

Email Marketing Powered by Mailchimp