Development of a photonic weed sensor for weed identification and spraying
The Weed Sensor and Spraying System project involves a partnership between ECU and Photonic Detection Systems Pty Ltd and seeks to develop a novel advanced photonic weed sensor prototype that captures and analyses spectral data for accurate discrimination and detection of weeds and crops.
No commercial products are available today to discriminate accurately between weeds and surrounding crops ("green from green"). The prototype enables the spraying of only weeds in situations including post-emergent crop growth.
ECU researchers integrate electronics, photonics and nano-engineered structures to realise multi-wavelength laser units of overlapped beams and aligned polarisaton states, as well as novel optical cavities for generating laser spots of spacing as small as 1.5cm. The researchers use a cost-effective, fast, integrated sensor array, (instead of discrete detectors) to simultaneously detect laser spot intensities, thus providing a much-enhanced spatial resolution. Researchers have also developed an efficient decision-making algorithm for real-time plant discrimination and weed control, making the system able to both detect specific weeds and spray at speeds up to 24km/h.
The outcome of this research will be a cost-effective weed management solution that significantly improves the profitability in agricultural and other sectors. This includes cost savings on herbicides (estimated at about 80 per cent), improvements in crop yields (about 30 per cent), measures against herbicide resistance, reduction in water use and major environmental benefits.
Key ECU staff
Professor Kamal Alameh, Dr Sreten Askraba, Dr Arie Paap.
Photonic Detections Systems Pty Ltd.