New weed-sensing technology being developed by the National Centre for Engineering in Agriculture (NCEA) in Australia may prove to be a breakthrough in weed control, offering automated identification of weeds in crops such as grain, cotton and sugar cane. The species-specific weed-sensing camera and spraying technology is still several years away from commercialisation, but promises to reduce both herbicide usage and resistance.
“Precision weed-sensing gear uses machine vision to differentiate between crops and weeds, and even between weed species. This is difficult when different weed species grow together and occlude each other,” says NCEA researcher Dr Cheryl McCarthy. Traditionally, farmers control weeds by applying herbicides such as glyphosate to the entire land.
However, as McCarthy points out, indiscriminate application of glyphosate to weeds on cotton and grain farms has led to glyphosate resistance and the appearance of hard-to-kill species such as fleabane and feathertop Rhodes grass. The cotton industry faces the added challenge of self-seeded Roundup Ready cotton, which grows as a volunteer in fallow lands and competes with other crops in minimum and no-till farming systems.
And in the Australian sugar industry, selective herbicides are costly, while run-off is environmentally damaging, particularly to the Great Barrier Reef. Another problem is that commercially-available technologies for spot-spraying weeds are not designed to operate in lands where weeds and crop coincide. “The technologies target any vegetation on a soil or stubble background, thereby limiting their application to only the inter-row or pre-emergent and fallow lands,” explains McCarthy.
Instant detection
In a bid to meet these challenges, the NCEA set out to develop a machine that could identify problem weeds. This meant dealing with difficulties such as inconsistent light and interference from ground cover. In the NCEA’s prototype, cameras and LED lights are mounted on the rear of a tractor to capture images of the crop row. These are then analysed with algorithms to gain a detailed analysis of the colour, shape, texture and height of plants. When a weed is detected, solenoid-controlled spray nozzles are triggered to apply the herbicide.
The cameras are able to detect broadleaf and grass species automatically. McCarthy and PhD student and engineer Steven Rees are currently developing a proof-of- concept weed identification and spot-spraying system for the cotton, grain and sugar industries. In the cotton and grains project, a sensing apparatus was fitted to a 3m boom; in the case of sugar cane, a single-row weed spot-sprayer was developed to target Guinea grass.
“The technology offers greater control of weed management by enabling selective spot-spraying with the appropriate herbicides,” explains McCarthy. “In addition, it is hoped that across-season in-field mapping will allow targeted application of pre-emergent herbicides to prevent an infestation from returning the following season.” Future developments should also allow for accurate analysis of ‘before’ and ‘after’ weed numbers.
Evaluations of the technology by spray operators have commenced on several farms. The NCEA also recently launched a weed identification and spot-spraying project for the Australian pyrethrum industry, which produces up to 80% of the world’s natural pyrethrum. “Botanical Resources Australia is collaborating with the NCEA in the pyrethrum industry and it will be evaluating the NCEA’s weed-sensing system for pyrethrum under commercial conditions over the next three years,” says McCarthy.
The precision weed-sensing technology has the potential to maximise yield and reduce chemical usage and run-off, while reinforcing the benefits of minimum-till farming. “The technology presents an alternative option for hard-to-kill weeds that might otherwise be controlled by tillage operations,” explains McCarthy. “The next few years will be very exciting as the research develops.”
Boost to phosphine’s effectiveness
Australian scientists discover the gene that governs an insect’s resistance to this fumigant.
Australian researchers have identified the gene responsible for an insect’s resistance to phosphine, a major fumigant used by the grain industry to ensure insect-free grain.University of Queensland scientist David Schlipalius believes that prolonging the effective life and availability of phosphine would be in the interests of the global grain industry.
“Phosphine fumigation is the most effective means of controlling pest infestations during grain storage,” he explains. “It is by far the most common treatment for grain and other commodities worldwide because it’s cheap, can be used on many commodities and doesn’t leave toxic residues.” The discovery, made by scientists from the university and the Queensland Department of Agriculture, will allow the researchers to better understand how resistance develops, providing insights into new fumigants or treatments that may block the resistance.
Schlipalius says that problem insects can now be detected even before an outbreak of resistance occurs. According to university colleague Paul Ebert, resistance occurs when genetic variants among the original insects survive phosphine fumigation and reproduce, generating resistant offspring. “The discovery of the resistance gene is the first step in identifying ways in which the resistant insects are vulnerable – their Achilles heel, so to speak,” he says.
Researcher Pat Collins says that the discovery is important because in some countries, losses in storage from grain insect pests alone can be as high as 20%. “Losses in India amount to about 20 million tons of grain per year – equal to more than half of Australia’s total production,” he explains. “Phosphine resistance would be a disaster for countries relying on it heavily, and that includes Australia, where 80% of our grain is treated by phosphine fumigation. It would mean the introduction of less effective and more expensive treatments, significantly increasing the price of grain and reducing its availability.”
Email [email protected]
