The new weed detection technology will help
farmers reduce their herbicide costs and
also benefit the environment.
Image: iStockphoto

A group of researchers at Edith Cowan University are developing a weed identification system for the farming industry that aims to specifically detect and target weeds, eliminating the need for blanket spraying.

Developed at ECU’s Electron Science Research Institute (ESRI) with industry partner Photonic Detection Systems, the high speed weed detection system, which has attracted international interest, will provide economic and environmental benefits to Australia.

“This system is based on detecting the weeds and spraying only the area where the weeds are located,” Dr Sreten Askraba, a senior research fellow at the ESRI, says.

“This will dramatically reduce the usage of herbicides to improve the quality of crops and also instantly reduce the impact on the environment.”

The system works by transmitting three lasers sources, which strike the same spot on the plant, collecting its optical properties. This reflected light is then processed by a control unit which determines whether the plant is crop or a weed. The spraying unit is activated if the plant is identified as a weed.

Director of the ESRI, Professor Kamal Alameh, says this weed detection system is unique in that it is able to discriminate between green weeds, green crops and brown soil, unlike other systems which can only discriminate between green weeds and brown soil.

“We are able to discriminate green from green and this is the world’s first device that can discriminate green from green, so weeds from crops,” he says.

The whole system, which incorporates laser spectroscopy, optics and thin-film technology, will be placed in one box and function as a self-sufficient unit.

“The plan is to assemble 30-60 of these boxes on the boom bar of the farming vehicle,” says Dr Askraba.

“The system is designed in such a way that the whole operation, which is a sequential operation of the lasers and data collection, will be computed in 1 millisecond and that corresponds to a farming vehicle of ~ 22 km/hr.”

Testing for first field prototype for the sugar and cotton industries will occur in November this year. The group is currently investigating whether the system is capable of detecting the notorious skeleton weed.