QUT researchers are using artificial intelligence to detect weeds that are replacing native plants in coastal areas.
Broadleaf pepper is a weedy tree that grows in the bushland foreshore areas along the coastal strip between Points Cartwright and Wurtula.
Native to South America and introduced as a garden ornamental plant, the tree can grow up to 10 m tall, fleeing gardens and invading coastal areas. Its fruits are spread by birds and animals, and the plant contains toxic resins that can affect human and animal health.
But the biggest concern is that it will suffocate and replace native plants, throwing away our precious sand dune biodiversity.
Peter Cox, Sunshine Coast Environmental and Habitability Council, said the trial combined three innovative techniques for detecting invasive weed trees.
“The first step that was completed was using a drone to capture images,” said Cr Cox.
“In the second phase, we will partner with QUT, which trains artificial intelligence to recognize weed trees.
“And finally, multispectral mapping uses sensors to map the test area.
“This will give us more information to help us manage the area so we can continue to protect and enhance its biodiversity.
“This particular location between Point Cartwright and Wurtula is pretty tricky because unexploded ordnance may still be in the area, so soil can’t be disturbed in some areas.
“The area was fired upon during World War II as part of artillery training, and the shells did not detonate when they landed as designed.
“We wanted to see if we could use the information we collected to create a detailed high-resolution map of where broad-leaf peppers are located within our reserves, in addition to the weed management already in place in the area. Try solutions to remove weed trees.
Cr Cox said the technology is being trialled on a small scale to test opportunities and constraints to see if it can be applied at scale or if it can identify other weeds.
QUT researcher Dr. Fernando Vanegas Alvarez says advances in artificial intelligence, drone technology, and image sensors have the potential to greatly benefit environmental management and ecological assessment.
“At QUT, we are transferring knowledge and expertise from remote sensing, machine learning and deep learning research to enable detection, assessment and mapping of vegetation over large areas,” he said.
“In particular, we are developing models to detect exotic species such as broadleaf peppers and assess the health status of native species such as pandanus trees along the Sunshine Coast.”
Funded by the Environmental Levy, the trial is part of the Council’s Invasive Weeds Project, which explores alternative weed management practices that can be used to control invasive plants throughout the region.
