We are seeking a motivated student for an MSc project. This candidate will join our research group and undertake a project that aims to develop a computer algorithm to detect marine mammals in remotely captured images.
This is a marine mammal related project but we are seeking a person skilled in programming and/or image analysis.
Project
summary:
The conservation and management of many marine mammal
populations relies on
accurate and precise estimates of their abundance,
distribution and habitat use
by conducting aerial surveys. An Australian federally funded
project is
currently being undertaken at Murdoch University to
determine whether Unmanned
Aerial Vehicles (UAVs) can replace manned aircraft.
The volume of images collected during an aerial survey, most
of which do not
contain marine mammal sightings, presents a processing
challenge post-survey.
During a previous project, images were collected at a rate
of one per second,
producing 3600 images per hour surveyed which then had to be
manually reviewed.
This processing is extremely time-consuming and fatigue
undoubtedly leads to
missed sightings.
Our research project aims to develop algorithms for
automatic detection and
classification of marine mammals within each image as a
feasible alternative to
manual analysis. A proof of concept algorithm has been
developed as a first
step. The approach requires further development to reduce
the
false-positive detections and decrease missed detections to
a rate equal to or
better than human.
Project collaborators:
Murdoch University Cetacean Research Unit (MUCRU)
Australian Research Centre for
Aerospace
Automation (ARCAA)
Requirements:
Academic background – Bachelors
in
Electrical/IT Engineering, or a related field with
preference to candidates
with expertise in the following areas:
1. Image processing.
2. Software programming:
MATLAB, C/C++, C#
If you are interested in this position, please contact
Amanda Hodgson (a.hodgson@murdoch.edu.au)
for
further details.