Researchers at RMIT have recently developed technology that is able to detect the types of trees and plants that are more likely to cause powerline faults, potentially sparking bushfires.
Two scientists, Dr. Yidan Shang and Nan Li developed the technology which can be used by power companies or manufacturers to detect and alert operational managers to the presence of vegetation on a high-voltage bare-wire powerline. The research won the first prize of the Vegetation Detection Challenge organised by the Department of Environment, Land, Water & Planning.
The Challenge was established by the PowerLine Bushfire Safety Program.
“Using the large amount of data collected through the Vegetation Conduction Ignition Testing project, we found patterns which we could model with,” said Dr. Shang, Research Fellow in the School of Engineering.
“We developed a machine learning technique based on the Convolutional Neural Network (CNN) architecture that can successfully identify what species of vegetation could pose a fire risk in touching the powerline.”
Associate Professor Sherman C.P. Cheung from the School of Engineering congratulated Dr. Yidan Shang and Nan Li.
“Their research is just one example of RMIT working with environment or private organisations to develop solutions to help the community,” he said.
Shang and Li received $ 10,000 in prize money as well as the support to further develop their research.
“I’m excited that we won the first prize because the government will support us to continue working on this project,” said Nan Li.
“The development of this research will continue to make communities safer from bushfires. It could potentially be implemented not only around Victoria but also around Australia or even the world.”