PhD Scholarships for a Robot Pruning Project.
University of Canterbury, Christchurch, NZ.
Two scholarships are available in the area of Computer Vision (Department of Computer Science) and a third scholarship in Robot Control (Department of Mechatronics Engineering).
These posts are part of the Ministry of Science and Innovation/NZ Wine growers funded project 'Vision-Based Automated Pruning' which aims to develop a robot system for pruning plants. As members of the project team the successful applicants will have the opportunity to combine leading computer vision research with a challenging real-world application.
Description.
The PhD student will work closely with a team of mathematicians, engineers, and computer scientists on a commercially-oriented project. The project provides an excellent opportunity to develop skills which are in high demand in both academia and industry, including scientific software engineering and the application of mathematics, statistics, and the latest algorithms. The ideal candidate will have a solid background in either mathematics or computer science, will have the ability to learn a range of new skills and keep up-to-date with the latest research, and will have the enthusiasm and motivation necessary to conduct innovative research in an exciting and fast-moving field of study.
Three three-year PhD scholarships are available with a stipend of NZ$30,000 per annum and the PhD tuition fees of $4,000 per annum. Additional support for equipment, travel, and conference attendance will also be available through the project. The three key areas of research are:
- (Computer vision) Recovery of three-dimensional structure of vines from vision and range-based cameras. We seek to establish a mm-accurate colour depth map of the vines through a combination of multi-camera sensor fusion.
- (Computer vision) Construction of detailed 3D models of the vines from this three-dimensional imagery. This will include models of the branching structure, orientation, and bud locations of the vines, as well as analysis of the images for signs of disease or damage to the vines. AI rules are applied to this information to estimate pruning cut locations.
- (Robotics/mechatronics) Path planning and control for multiple robot arms attached to a moving platform. The path planning and control will dynamically integrate the output of the computer vision system, odometry information, and knowledge of the kinematics of the robot arms. The system will be tested both with an existing 6DOF robot arm and in simulation.
Eligibility.
Applicants should have an Honours degree or higher in Computer Science, Applied Mathematics, Mechatronics or a related discipline and should possess:
- Strong ability in mathematical modelling and analysis
- Computer programming for scientific applications, preferably with knowledge of C/C++
- Experience/knowledge in one or more of the following areas:
- Reconstruction of depth from stereo images
- Sensor fusion
- Concurrent and real time image processing
- Numerical optimisation
- Shape analysis and statistical modelling
- Image classification and texture analysis
- Robot path planning and control
