A prototype drone equipped with a camera, thermal imaging software and carousel which can carry 10kg of baits, has been put to the test by pastoralists looking to identify and control feral pests in remote Western Australia.
The drone, which is remotely controlled from an office desktop, can carry 72 baits and has a flight time of 1 hour 20 minutes at maximum speed of 140km/hr.
3D printing technology was used to create the bait carousel with sausage bait as the model.
A second petrol/electric hybrid model with a four to six-hour flight time and 1100 bait carousel is also being developed.
The drone mounted species recognition system is part of a project underway in the Murchison region of WA’s Southern Rangelands, funded by the Australian Government’s Agricultural Innovation Hubs Program, with support from the Southern Rangelands Pastoral Alliance, through the South-West WA Drought Resilience Adoption and Innovation Hub.
SRPA Executive Officer Margie Weir said the project is aims to provide early warning of feral animal incursion into pastoral properties, and to contribute to a data base of feral animal movements and their damage to the environment and interaction with other livestock.
The idea for the drone was conceived by Murchison regional pastoralists Mayne and Leanne Jenour and RC helicopter (drone) pilot Nigel Brown of Autonomous Technology, Perth.
The Jenours run 1200 Santa Gertrudis/Droughtmaster breeders and 1800 Damara/Dorper ewes across five pastoral stations in the Murchison region of WA totalling around two million acres.
Nigel has been developing drone technology and sensors for the past 18 years.
“We program a flight path across the rugged, inaccessible country and drop baits into there,” Mayne said.
A grant obtained by the Southern Rangelands Pastoral Alliance is enabling the pair to add species recognition software developed by UNE, the NSW Government and Centre for Invasive Species Solutions, called the Wild Dog Alert Node, for large feral herbivores including donkeys, camels and wild horses.
The solar powered Wild Dog Alert Node incorporates technology unavailable on commercial camera traps, and contains a small computer which processes the images captured by a 360 degree camera and sensor system.
Onboard artificial intelligence software detects the presence of a dog and sends a text alert message via satellite to the Wild Dog Alert Cloud service on the user’s app or email.
“The idea is to send the drone out to the station’s peripheral areas where there are incursions of large feral herbivores. It will use thermal imaging (body heat signal) to locate the animal and species recognition software to identify the feral pessts and then alert the pastoralists,” Mayne said.
“This will give people an early warning of what is coming – once the animals get to a water point it is too late as they exclude the cattle from drinking and smash the infrastructure.
“It can be looking for and baiting wild dogs at the same time as monitoring water troughs and windmills during flights.
“At the moment we have line of sight communication (20-30km) to the base station but do have the facility to link with a cellular network. We are looking at utilising Elon Musk’s Starlink system with one node connection to the homestead and a series of towers to create a WiFi mesh across the station.”
“The drone will beam information in and out of the WiFi mesh to the operator.”
Nigel said it was about designing ease of loading, dispensing and dispersal. “It had to be lightweight and simple to use, and we came up with a unique concept of single drop delivery where the drone can deliver a single drop of baits, a large bait or three smaller baits,” he said.
“When the release mechanism is triggered, the GPS location is recorded on the drone.
“The helicopter drone is a stable platform, operates in adverse weather conditions and has long endurance for a drone.”
Mayne and Nigel are refining the algorithms, software, machine capacity and bait carousel size before commercialisation.
For more information on this story on the wild dog plan website click here