Weedetect® – Bilberry

Traditionally, herbicides are applied to a paddock in a blanket spray in an effort to reduce weed numbers, and ensure that there are minimal survivors which decreases the chance of seed set. In fallow or summer spraying situations, this has historically been achieved through blanket spraying broad spectrum or non-selective herbicides to eliminate all plants present.

This can be done for many reasons, including:

• Conservation of moisture for the following crop
• Reduce the green bridge for potential disease carryover
• Machinery trafficability and less blockages at seeding time

Spot spraying is the alternate way in which weeds in fallow and crop situations can be managed. This reduces the cost of herbicides, saves water, aids in the reduction of herbicide resistance by being able to deploy stronger chemicals on a smaller scale and saves time due to less filling up of boomsprays.

Historically, fallow spot spraying has been achieved through the use of chlorophyll sensors to identify weeds on bare paddocks, or green plant matter on soil (Green on Brown). This technology has been available for many years with good results, and has reduced the herbicide use of non-crop situations up to 90%.

• Measure the reflectance of green plants (using near infrared wavelengths) which triggers the weed to be sprayed
• Can be used day and night as they have their own light source
• Can reduce herbicide application by 80 to 90% (depending on density of weeds)
• By reducing area sprayed it also reduces number of tank fills required (fewer litres of spray mix per ha)
• Boom height control is critical to ensure adequate coverage of weeds
• Coverage is affected by strong winds which moves the spray pattern away from the targeted weed

*Refer to below list of commercial applications

• Advanced AI capabilities – No longer limited to chlorophyll sensors; cameras can now distinguish weeds from crops (“Green on Green”)* technology
• Industry-wide development – Start-ups, corporations, and universities are creating similar systems using AI with RGB or hyperspectral cameras

*Refer to below list of commercial applications.

• Deep learning technology – Uses computer vision to recognize objects in complex images with high variability
• Three key enablers:
  • Graphics processing units (GPUs) for computing power
  • Big data sets for training
  • Complex neural network algorithms
• Brain-inspired design – Uses deep neural networks that mimic human brain function with supervised or unsupervised learning

• Camera systems are affected by high stubble loads, crop shading, and canopy closure – a limitation shared by all optical sprayers (GoB or GoG)
• Camera systems will never achieve 100% hit rate – a two spray strategy is highly recommended
• If cameras cannot see weeds, they cannot spray them
• Night spraying is green on brown compatible only
• Setting realistic expectations with potential users is paramount
• Success comes from happy users whose expectations are met – underpromise and overdeliver

Two Spray Strategy Benefits

• A two-spray approach ensures minimal weed survivors by targeting later germinating weeds
• Second spray timing catches escapees from the first application, significantly improving overall control

Hit Rate Calculations

• Individual spray hit rate of 80% = 96% total control when combined
• Individual spray hit rate of 90% = 99% total control when combined
• Individual spray hit rate of 60% = 84% total control when combined

System Integration

• Bilberry camera system is an additional tool for integrated weed management
• Works alongside double knocks, harvest weed seed management, and other Weedsmart ‘Big 6’ components
• Confident the system will achieve 80%+ hit rates per application