The real-time plant identification and spaying system has been developed by Edith Cowan University researchers and Perth-based company Photonic Detection Systems.
The weed detector can distinguish between crops and weeds on-the-run and selectively spray herbicide where needed, reducing the need by 75 per cent.
Lasers of different frequencies are rapidly pulsed into two optical cavities, splitting the beams in 30-second intervals, with each emitting a precise volume of light to give maximum ground coverage.
A sensor captures and measures the spectral response, which is communicated to the central processing unit and compared with pre-programmed criteria.
Tracker lights indicate findings—green lights show vegetation, yellow show crops and red highlights weeds.
Shades are placed over the device to reduce ambient light, which can affect the return signal.
WA’s Chief Scientist Professor Lyn Beazley described the photonic weed detection system as a simple concept—but one with a lot of “boffin power” behind it.
“It uses innovative laser technology where the beam is reflected down to the crops and back again and, by looking at the quality, it tells us whether it’s a crop or a weed,” she said.
Believed to be a world first, PDS had almost ‘given up’ on the venture six years ago until it met with ECU Electron Research Institute Director Professor Kamal Alameh, WA’s 2007 Inventor of the Year.
“It’s proved to be a wonderful partnership and example of how industry and universities can work together,” PDS managing director John Rowe said.
“There are other companies that can distinguish between plants and non-vegetation—green from brown—but this is the only machine in the world capable of distinguishing the green of two separate plants.
“The benefits to farmers include less herbicide use, less crop stress, increased yields and environmental benefits.
“Agricultural has to move with the technical age to survive. Precision agriculture can micro-manage the paddock and this is the missing link.”
While a production model is yet to be developed (only a demonstration model exists), Mr Rowe said plans were under way to further improve the photonic weed detection system.
“We need to expand our knowledge of plant spectography,” he said “We can’t look at a database because it doesn’t exist yet.
“We also need to make the machine go faster and be able to detect faster. And we need to be smart and more sophisticated in the algorithms we’re using for detection.
Original Article Here
No comments:
Post a Comment