Edge computing platform for bird repellent system
Automated remote sensing technologies are increasingly being developed for agricultural applications. The huge amounts of data they provide to farmers is only useful when processed, a task that currently relies on cloud-based data centres. The remote locations and vast properties of farms also creates challenges in terms of access to electricity and bandwidth to support these operations.
This project aims to develop a solar-powered edge computing platform to address these problems, demonstrated with an automated pest bird repellent system. Damage to crops from birds is a huge problem for Australian farmers, but automated deterrents must be immediate to be effective. Using machine learning to detect birds with image processing and task execution handled on site in embedded edge computing platforms will speed up processing and enable responses in real-time.
Another goal of this project is to optimise methods of energy distribution and task prioritisation in a network powered by potentially intermittent solar energy. This platform could help increase reliability of solar energy networks in agriculture and other sectors.
Dr Adel Nadjaran Toosi