Any fisheries management regime depends upon an understanding of recreational fishing levels, and ensuring that the amount being fished is sustainable — but how to get that information? This was the topic of Jing (Jane) Zhao’s doctoral thesis in Computer Science at Unitec, a three-year project funded by the Ministry of Primary Industries (MPI) and done in partnership with the National Institute of Water and Atmospheric Research (NIWA).
Currently, NIWA monitors traffic at key boat ramps around New Zealand using CCTV cameras, but technicians then have to use that footage to manually count the number of boats, which is a time-consuming and costly process.
Using a discipline called Computer Vision, Jane looked to find a way to use computers to automatically monitor the traffic information provided by images.
While there were existing methods that could automatically count objects from video footage, says Jane, these were designed to monitor footage that typically had 25 images/second, while NIWA’s web cameras captured only one image/minute.
Using computational intelligence and image processing techniques, Jane developed a new object tracking algorithm that could potentially track boat movement at the lower frame rate.
After completing her thesis, “Computer Monitoring Trends in Recreational Fishing Effort” Jane worked for several months with NIWA to transfer knowledge and help the organisation find ways to use her research to improve existing processes.
Jane’s thesis was co-supervised by Paul Pang (professor of Computer Science at Unitec), Hossein Sarrafzadeh (professor of Computer Science at Unitec) and Bruce Hartill, who leads the research NIWA conducts on recreational fishing. “While we need to do more work in this area before we can use Jane’s methods to automatically monitor boat traffic around New Zealand, we’ve certainly learned a lot about Computer Vision methods from this research. It has improved our understanding for the issue and how we might resolve it,” says Bruce.
It is very unusual for MPI fisheries to fund a PhD, he adds, which reflects the need to find a way to better ways to understand the scale and impact of recreational fishing in New Zealand “It’s a very big problem, and Jane has certainly helped us get closer to developing a solution.”
Jane has published five papers as a result of her research, and has also been employed as a lecturer in Computer Science at Unitec. The project was often challenging and sometimes frustrating, she says, but she relished the chance to work on a real-world project aimed at helping solve a real-world problem. “That was motivating. Unitec is fantastic for having provided me with that opportunity.”
Find out more about studying Computer Science at Unitec: