Microspraying with UK’s SRC Tomv4 hits the target

During the past two years the team has worked on an Innovate UK Smart Farms project to develop a robot system for monitoring, analysing and treating weeds, pests and diseases in maize and wheat crops.

“Microspraying could be game changing for the industry,” says Ben Scott-Robinson, CEO of SRC. “Pressure is increasing from regulators, leaving farmers short of options. Precision spraying could enable a new generation of spot treatments, cut costs and significantly reduce the impact on the environment.”

_{Text continues below picture}

Significant milestone

This development marks a significant milestone for the company based in Salisbury in the county of Wiltshire in the UK. It is the first time it has integrated a third-party tool into its commercial Tomv4 robot.

This opens the way for it to play host to a range of other systems, enabling it to extend the platform’s same benefits of using small, light robots for treatments as well as for detection.

It does not, however, scan and spray in a single operation. It still uses its existing ‘See then Spray’ process, with the Tom robot first scanning and mapping every plant and weed to a ‘PerPlant’ resolution of 0.28mm/px.

This data is then analysed to produce a treatment map, which is used to control the on-board sprayer in a separate operation.

_{Text continues below picture}

Maintain the accuracy

Ray King, Technical Lead for R&D projects, explains this way it can maintain the accuracy required for microspraying to a resolution of 25cm x 25cm. In the future it could combine both operations, but this will require changing the algorithm and software to enable it to process the data on-board quickly enough to react in real time.

“The robot currently captures and retains all data captured in the field, whether or not it is useful. We process the data in the cloud, removing the useful information and producing the valuable insights at this stage. We have projects in place that aim to develop the robot to allow real-time processing of the data to remove any data that isn’t useful to the farmer.

“For this, I like to use the analogy of a combine harvester, ingesting all of the crop and processing it while moving through the field, retaining only the valuable grain and leaving behind the less valuable crop residue, except the crop in our case is data not wheat,” explains Mr King.

_{Text continues below picture}

Spot a single leaf weed

The Small Robot Company adds it has the first and still only system that can detect and map grass weeds at the single leaf stage on a field scale. Interestingly, SRC doesn’t use any special sensors to detect these single, thin weeds.

“If you can see it with your eyes in the field, then developing a system to also see it is not that difficult. We use standard RGB cameras – it’s the software that is important,” says Mr King.

While for scanning the robot can achieve a normal workrate of about 20 ha/day, it’s easy to increase the capacity to work in a large arable field by simply adding more robots, rather than expanding the size of the equipment.

When it comes to treatments the sprayer’s overall output will vary depending on the area of the infestation. Weeds usually occur in patches and the robot will not only treat just these specific areas, but also calculate the shortest route to cover the field.

_{Text continues below picture}

Future developments

Currently the mircospraying system is using on/off individual nozzle control. In future Mr King sees the use of Pulse Width Modulation to not only provide targeted applications, but also vary the dose rate to match the level of weed or pest infestation.

“These very precise weed maps can also be superimposed over the eventual crop’s yield map, which can be used to assess the impact and the efficacy of the treatment,” he says.

Moreover, they can also be integrated into creating nitrogen variable rate map, which relies on using the Green Area Index (GIA) to identify the differences.

“Now, however, we can ensure that high weed infestations (which will increase the GIA) don’t influence the calculations. This way we can ensure we can choose not to focus on feeding grass weed hot spots,” adds Mr King.