Read time: 4 minutes

Advances made in weed recognition technologies

Blanket spraying of herbicides in low weed density situations could become a thing of the past in Australia as new weed recognition technologies enable site-specific weed control in-crop.

Just as Australia led the way with the development and adoption of ‘green-on-brown’ weed detection and spot spraying in fallow situations, now Australian researchers are developing technologies that will deliver ‘green-on-green’ weed recognition and targeted control in-crop.

Imagine a machine that can identify one weed species from another and apply the best treatment to each weed, even in-crop. While expert human brains can make these differentiations and decisions relatively easily, training artificial intelligence technologies to do the same thing is challenging.

With investment from GRDC, a team of researchers led by Dr Michael Walsh, director weed research at the University of Sydney, have recently completed the pilot phase of crucial work that will underpin future developments for machine learning in weed recognition.

Dr Michael Walsh, director weed research at the University of Sydney, says the WeedAI image database will underpin future developments for machine learning in weed recognition.

Dr Walsh says there are several commercial interests developing machine learning technologies for site-specific weed control in Australia, but they all need access to a collection of relevant images to essentially ‘train’ computers in the development of weed recognition algorithms that can differentiate between crop and weed plants.

“We have set out to establish protocols for collecting and annotating images that will be stored in an open-source database that anyone with commercial or academic interests can contribute to and also use for future developments in this technology,” he says. “The pilot project has centred on collecting images and developing weed recognition algorithms to detect representative grass and weed species in wheat and chickpea crops.”

The WeedAI database currently contains thousands of images of annual ryegrass and turnip weed growing in chickpea and wheat crops. These images have been manually annotated and used to develop and test weed recognition algorithms for their accuracy in correctly identifying weeds growing in-crop.

“The images are all high quality, with annotation outlining the weed shown in the image and notes about the agricultural context, such as soil colour, location, crop type, and growth stages of the crop and or weed,” he says. “We are hoping to fast-track developments and take advantage of the machine learning technologies that have capability to accurately recognise and locate in-crop weeds to ultimately provide growers with the opportunity to specifically target these weeds with a range of weed control options.”

Machine learning offers the potential for high-level accuracy in weed recognition in-crop.

“We are hopeful that this will give growers access to a range of novel chemical and non-chemical weed control technologies that will add to the existing options available for in-crop weed control. This might include herbicides that are currently too expensive for blanket spray application.”

Dr Walsh says Australia is leading the way in developing weed recognition technologies for grain production systems and he believes the open-source database will reduce replication of effort and encourage technology companies to address more challenging scenarios, such as recognition of grass weeds in cereal crops.

Like the optical spray technology that brought tractor-mounted spot spraying to fallow management over 20 years ago, the green-on-green in-crop weed recognition systems in-crop will be used for site-specific weed control in situations where weed density is already quite low.

“At densities of less than one weed per 10 square metres, the area sprayed with herbicides would be 70 to 80 per cent less than when a blanket spray is applied,” says Dr Walsh. “The opportunities to introduce different herbicide modes of action or alternate methods of weed control such as targeted tillage or laser treatment can also be considered to reduce the risk of herbicide resistance.”

Ground speed is the enemy of real-time weed recognition systems, as accuracy increases considerably with speeds slower than those currently used for blanket spraying. With increasing computing processing speeds the expectation is that in-crop weed recognition systems will be accurate at 10 to 15 km/h. The introduction of autonomous platforms is reducing the need for higher speeds, and with a light source there will be the opportunity for round-the-clock operation of weed recognition equipped site-specific weed control systems.

A number of commercial companies are bringing in-crop spot spraying to market and will be on-hand at WeedSmart Week, Esperance to showcase their technology in mid-August. Ben White, Kondinin Group’s research manager will host the machinery session with spray and harvesting gear on display including Goldacres’ G6 Crop Cruiser series 2, weed detection technologies using drones, weed identifying cameras (green on green) and a range of harvest weed seed control options including impact mills from Seed Terminator, Redekop and iHSD (both hydraulic and belt-driven) and the Emar chaff deck. This flagship event always attracts growers keen to see how other farmers are keeping weed numbers low in different systems. Early bird registration is now open.

More resources

Related Articles

View all
Article
News

Investigate adverse experiences when using herbicides

A shuttle of glyphosate applied over the top of a Roundup Ready cotton crop was recently shown to also contain a damaging level of 2,4-D impurity, resulting in significant crop injury and yield loss.
The grower involved did not accept the suggestion that the crop damage was due to poor sprayer decontamination or spray drift from a fallow application of 2,4-D, and he was able to prove the problem was due to product impurity.
2,4-D herbicide injury in cotton after the crop was sprayed with glyphosate product contaminated with 2,4-D.
Other shuttles of the same batch may have been applied to fallow weeds where the residual 2,4-D in the glyphosate would have gone unnoticed. Full rate 2,4-D in glyphosate is known to compromise glyphosate efficacy, but studies of low-rate 2,4-D impurity in glyphosate could not be found.
Where can impurities come from?
While the agricultural chemical manufacture and supply chain in Australia is considered first-class and is highly regulated, there is an acceptance that the nil-impurity requirement for the manufacture of agricultural chemicals is unattainable in facilities that use multi-purpose equipment for synthesis, formulation and packaging of products.
Companies therefore apply their own quality assurance standards before releasing products for distribution and sale. If the level of risk posed by certain residual impurities in a product is underestimated, there is potential for instances of crop injury, pesticide residue in produce or poor performance of the product on the intended target weed, fungus or pest.
Mistakes can and do happen within the manufacturing process and chemical supply and distribution chain. To ensure that risks of contamination are minimised and that quality assurance protocols are followed carefully, it is important that any breaches or errors are identified quickly, reported and investigated.
Keep good records of each spray event, including batch numbers of applied product, to help identify the cause of adverse experiences with herbicides.
There are two important things to note: firstly, the current regulations specify that crop protection products must contain nil impurities (other than manufacturing impurities listed in the APVMA standard); and secondly, companies are required to recall product batches when contamination issues are identified. The Australian Pesticide and Veterinary Medicines Authority (APVMA) oversees a highly regulated system of registration, compliance and enforcement on crop protection products.
Assess potential application issues
When misapplication (wrong product applied, incorrect mixing, contaminated product etc) occurs, symptoms of affected plants are usually uniform throughout the treated area. It is often suggested that poor application technique or poor sprayer decontamination is the reason for crop injury or poor weed control results – suggesting a grower ‘own-goal’. Such potential errors must be considered, but if best practice spray techniques and spray rig decontamination procedures have been followed, product impurity should also be considered and investigated.
The chemistry of the product will determine the risk of residues being held within the tank and spray lines of the application rig. This is why there are differences in the sprayer hygiene requirements after using particular products.
Most modern spray rigs have impervious rubber and plastic, or stainless steel components, drastically reducing the risk of chemical absorption and subsequent extraction. Residues on the rubber surfaces are the main concern, and all registered cleaners will physically remove residues when used as directed, but cracked rubber components can present a contamination risk. All filters/strainers must be cleaned and all actuators and taps musts be cycled as the cleaner is run through the spray boom and tank loading system, agitators and tank.
Crop injury or poor weed control that is associated with just one sprayer tank load would suggest sprayer contamination. Effects from contaminated tanks are usually worse at the beginning of the spray run, with damage diminishing with spraying and tank reloading. The field pattern can provide clues to the sprayer filling routine in the field where the crop damage occurred.
The other major reason commonly cited for crop injury in spray drift. Although there is always some small amount of drift when agricultural chemicals are sprayed from a ground rig, the amount is down to ‘virtually safe’ levels within a few tens of metres. If the conditions are very windy, or the boom is too high, or the droplet size too small, spray could drift a few hundred metres from the application ground rig.
Spray droplets may travel a few feet to several kilometres from the targeted area, depending on weather conditions and spray application; but the potential for drift damage decreases with distance because droplets are deposited or become diluted in the atmosphere. The pattern of injury is normally seen most prominently on the section of the field closest to the sprayer that generated the spray drift, and decreases across the field.
During inversion conditions, a similar amount of product is subject to drift, but the drifting product will not dilute as much in the air, so concentrations at specific locations can be higher than expected in non-inversion conditions.
What to do if your crop is damaged or weeds don’t die as expected?
Along with several other possible causes, unintended application of contaminated product should be considered as a potential explanation for crop injury or poor weed control.
Keep in mind that if product impurity is the problem, it is most likely due to a low-dose effect that may be difficult to diagnose or may take longer to express in the target weeds or susceptible crops.
Finding the cause of an ‘adverse experience’ with herbicide is one of the most important reasons to keep accurate and detailed spray records.
If a problem occurs:

Take detailed, time-stamped photographs of the crop or weeds and record everything you know about the crop or fallow management, weather conditions in the weeks prior to the damage being seen, spray history of the field etc. If possible, geotag the photos so they can be easily associated with the correct field.
Record the relevant batch numbers of the chemicals used, which can be checked against the retention samples at the factory if necessary. Collect samples from drums of product used prior to the injury being observed (up to 14 days prior to symptoms being obvious). When you take samples, make sure there are witnesses who can vouch for the voracity of the evidence you have collected. Testing for one impurity (e.g. 2,4-D in glyphosate) costs less than $500 per sample.
Document the injury over time. For example, injury in cotton from low rates of 2,4-D will grow out in two weeks, but injury from higher rates, could last three to four weeks and are the most likely to result in yield loss. Similarly with weeds although the impact may be more difficult to document.
Mark out the affected area in the field to help assess crop yield loss at the end of the season. Note the pattern and intensity of the problem across the field.
Eliminate as many possible causes as you can. Re-assess the application technique and equipment, consider the pattern of damage in the field, look at the weather conditions for the relevant period of time and so on.
Test for herbicide resistance in weeds.
Report the crop damage or poor weed control. The APVMA administers the Adverse Experience Reporting Program, which allows anyone to report a problem with an agricultural chemical, including crop and plant damage, for example, plant death, severe stunting or significant yield loss. This is also the way to report poor weed control outcomes.

The APVMA acknowledges there is likely under-reporting of adverse experiences. The magnitude of under-reporting is unknown and provides limitations in quantifying product risk.
Investigations of spray drift are conducted by the relevant state government body, for example: NSW EPA (call Environment Line: 131-555), Biosecurity Queensland (call 132-523) and Chemical Standards Officer (Victoria) (call 03 5430 4463). Industry organisations will also support growers impacted by chemical damage to crops.
If the damage is due to factors other than spray drift, the affected party will need to take legal action and seek compensation themselves.
Related resources
Is poor weed control due to herbicide resistance?

Article
News

What to expect at WeedSmart Week 2021

Big 6 at WeedSmart Week 2021 – Double knock to protect glyphosate
The WeedSmart Forum is set for Tuesday 17 August, 2021 at the Civic Centre in Esperance WA. The program features growers, agronomists and researchers discussing ways to use the BIG 6 to beat crop weeds. You can register for the 3-day WeedSmart Week event here.
Greg Warren from Farm and General in Esperance will be sharing his thoughts on the control of weeds like summer-germinating ryegrass, marshmallow, fleabane and portulaca.
He says the growers around Esperance are tackling glyphosate resistance in annual ryegrass, along with brome and barley grass and other emerging weeds using a range of integrated control tactics. The double knock plays a key role in preserving glyphosate (and soil moisture) and providing a clean seed bed for planting crops.

 
Big 6 at WeedSmart Week 2021 – Increase crop competition
WeedSmart Week 2021 is set for Tuesday 17 to Thursday 19 August, 2021 in and around Esperance WA. The last two days feature local growers hosting visits to their farms and discussing how they use the BIG 6 to beat crop weeds. You can register for the 3-day WeedSmart Week event here.
One of the farms hosting a visit during WeedSmart Week is Warrakirri’s 12,800 cropping operation at Condingup. Farm manager Con Murphy has implemented a variety of tactics to combat their main weeds – annual ryegrass and wild radish. Since 2015 the farm has undergone an intensive soil amelioration program to improve the drainage and ameliorate the sandy soils across the farm.
Con says the benefits have been seen in better germination and establishment that sets their cereal, pulse and canola crops up to compete strongly with weeds. There is also a benefit at the end of the season where rain in August or September enters the soil profile without causing waterlogging, and providing a better finish for their crops.
Since 2016-17 about 80% of the farm has been ripped and a portion has been ripped 2 or 3 times because the sandy soils tend to slump after substantial rainfall events, recreating the hardpan.
Con will be showing the WeedSmart tour group how their ripping, drainage, liming and spading program has helped grow more crop and less weeds!
Listen to the podcast with Warrakirri’s Con Murphy talking about the impact of improved drainage on crop competition

Big 6 at WeedSmart Week 2021 – Implement harvest weed seed control
The WeedSmart Week machinery display is set for Wednesday 18 August, 2021 at Dave Campbell’s shed near Esperance WA. The 3-day WeedSmart Week program features growers, agronomists and researchers discussing ways to use the BIG 6 to beat crop weeds. You can register for the 3-day WeedSmart Week event here.
We’ve saved the harvest weed seed control discussion for the machinery session on Wednesday 18 August. Ben White from Kondinin Group will host the machinery session with spray and harvesting gear on display including impact mills from Seed Terminator, Redekop and iHSD (both hydraulic and belt-driven), Emar chaff deck, and spray technologies including Goldacres’ G6 Crop Cruiser series 2, and weed detection technologies using drones and weed identifying cameras (green on green).
Ben White, Kondinin Group (Photo: Melissa Powell, courtesy of GRDC)
Growers doing the WeedSmart Big 6
WeedSmart Week 2021 is set for Tuesday 17 to Thursday 19 August, 2021 in and around Esperance WA. The last two days feature local growers hosting visits to their farms and discussing how they use the BIG 6 to beat crop weeds. You can register for the 3-day WeedSmart Week event here.
One of the growers who will open up their farm for a visit is Adrian Perks who farms at Condingup, 70 km north-east of Esperance. Adrian runs a continuous cropping program on his 4300 ha property, growing canola, wheat, barley, faba beans and lupins. This diverse rotation allows him to mix and rotate both chemical and non-chemical weed control tactics. Over half of Adrian’s farm is sandplain, on which he has implemented a soil amelioration program to address non-wetting to increase the competitiveness of his crops. He currently uses chaff decks for harvest weed seed control and is introducing an impact mill this season. Adrian monitors the tramtracks for weed growth and if he feels the weed pressure is too high, he uses a shielded sprayer to reduce seed set. The bus tour will include four farm visits and a machinery display.
Listen to Adrian on the Regional Update podcast.
Adrian Perkins, Condingup WA
 

Article
News

WeedSmart agronomist set to tackle high rainfall zone weeds

Every locality has its own spectrum of weeds, and growers face different opportunities and challenges regarding the control tactics they can employ.
The WeedSmart Big 6 approach is a practical way to ensure that an integrated weed management program is put in place that disrupts weed seed production and the evolution of herbicide resistance.
Commencing in January 2021, Jana Dixon has joined the WeedSmart team of extension agronomists, with a focus on applying the Big 6 to manage weeds in the high rainfall cropping systems of southern Australia – from Esperance in WA to south-eastern SA, Tasmania and south-western Victoria.
Jana will add to the dedicated and experienced extension agronomists on the WeedSmart team with Peter Newman in the Western region, Chris Davey in the South, Greg and Kirrily Condon in the East and Paul McIntosh in the North.
Jana Dixon has joined the WeedSmart team of extension agronomists, with a focus on applying the Big 6 to manage weeds in the high rainfall cropping systems of southern Australia – from Esperance in WA to south-eastern SA, Tasmania and south-western Victoria.
Jana hails from the Mid North of SA, and began working at Pinion Advisory (previously Rural Directions) while she was studying agriculture at the University of Adelaide. She has been employed full-time at Pinion Advisory since January 2019 as an agribusiness consultant, based in Clare, and spends most of her time delivering agronomy and farm business advice to clients from a wide range of cropping regions in South Australia.
Pinion Advisory is a foundation WeedSmart sponsor and Jana has been involved in two WeedSmart Week events already – the first as a participant and grower group organiser at the Horsham event in 2019 and then as the local organiser for WeedSmart Week 2020 in Clare.
In welcoming her to the WeedSmart team, program manager Lisa Mayer says Jana brings energy, commitment and insight to deliver communications focussed on the southern region’s high rainfall regions.
“Growers in the southern high rainfall zones are facing some serious issues with herbicide resistance influencing their farming decisions,” says Ms Mayer. “Jana will be engaging with agronomists, growers and researchers in each of the distinct high rainfall zones to understand the complexities and look for practical ways to apply the WeedSmart Big 6 in various cropping scenarios.”
“We plan to deliver WeedSmart Week in Esperance, part of Western Australia’s high rainfall cropping zone, in August 2021 and Jana will play a key role in the planning and delivering of our annual 3-day flagship event.”
Jana says her experience with the WeedSmart program has been very positive and she has been particularly impressed with the support the program has from all sectors of the grains industry.
Newly appointed WeedSmart extension agronomist, Jana Dixon (green cap) leading discussions with farm visit host, Ben Marshman, Owen SA, and growers and agronomists attending WeedSmart Week 2020 in Clare.
“I have spoken to many growers and agronomists who have found real value in the information that the WeedSmart program delivers,” she says. “For many it is as much about considering another operator’s philosophy on dealing with weeds, and taking a fresh look at their own systems, rather than just learning about a new tactic or the traits of a new herbicide in isolation from the big picture.”
She says the high calibre of industry people who contribute their time and expertise to the program is testament to the value WeedSmart has to agribusiness, growers, agronomists and researchers alike.
In taking on the responsibility for delivering information tailored for the high rainfall zones Jana says she is pleased to have an extensive network of contacts through Pinion Advisory, with offices in a number of high rainfall areas to provide easy access to local agronomists and growers. She is also aware that there are major differences in weed spectrums and farming systems in each high rainfall zone and plans to take full advantage of the opportunity this role presents to expand her understanding of different approaches to weed management.
“The long and favourable growing season and the associated prolonged periods of weed germination, is a key factor that I see potentially impacting on a grower’s weed management strategies in these regions,” she says. “On the other hand, access to highly diverse rotations and a focus on crop competition are two strategies that can play an important role in achieving excellent weed management in these regions.”
“I am keen to engage with anyone working and farming in the high rainfall zones to build my knowledge and understanding,” she says. “And to create opportunities to develop and extend the WeedSmart Big 6 strategies, both herbicide and non-herbicide, that work in each area and in different situations.”
WeedSmart is the industry voice delivering science-backed weed control solutions with support from the Grains Research and Development Corporation (GRDC), major herbicide, machinery and seed companies, and university and government research partners, all of whom have a stake in sustainable farming systems.
You an follow Jana on Twitter and keep up to date with the HRZ here.

Subscribe to the WeedSmart Newsletter