Tech Bytes #6
Pretty fly for an AI spy
Why is it important?
Sam Watson Jones Precision AI is probably the company that we admire the most out there in the market at the moment. What they have achieved in terms of image capture at speed and scale is remarkable, and it has inspired us to significantly push our own technologies to increase our own operating parameters with our ground based robot. The vision of spraying from drones is certainly interesting, but it is definitely not without challenges. It will be intriguing to see how their proposed spraying drone differs in operating parameters to their scanning drone, and also how quickly the legislation moves to support this business model. We think these guys are great though and we’d love to work more closely with them in the future.
What is it?
Canadian start-up Precision AI has launched its new drone concept at the Ag Motion event in Western Canada, earning the company the Farmers Award for Innovation in Environmental Sustainability. A hybrid VTOL (vertical take-off and landing plane), designed to be deployed in hives, the autonomous drone promises stable, drift-free spraying, staying airborne for over two hours at speeds in excess of 45mph. Precision AI, that raised a record CAN$20M (£13M) in seed funding last year, claims the on-board cameras have a resolution of 0.05mm and can deliver green-on-green spot spraying, differentiating instantly weeds within a range of broad-acre crops.
Satellite tech takes the carbon credit
Why is it important?
Sam Watson Jones The carbon market is struggling to receive take up by farmers nervous about the lack of an agreed standard. I think the combination of drones and satellites working in unison to measure the characteristics and farming practices in fields, and thereby interpret something about the carbon sequestration in the soil is interesting, but I am still somewhat sceptical. I think we should be getting more granular with our soil measurements, and that means ground based, autonomous machines. This is certainly not without challenges, but without trusted, detailed datasets truly measuring the variability in soils, there will always be a handbrake on this market.
What is it?
Carbon farming company Agreena has acquired Hummingbird Technologies to deliver its monitoring, reporting and verification (MRV). The aim is that Hummingbird’s data and automation capabilities will add integrity and trustworthiness to soil carbon certificates issued by Agreena. Low-orbit satellite imagery, combined with ground-truth data, is used to automatically recognise and report on-farm regenerative agriculture practices such as cover crops, tillage, and crop rotations to track green biomass and continuously monitor field activities, says Hummingbird. Denmark-based Agreena now operates in 13 countries across Europe and was one of the first platforms to trade carbon certificates for farmers.
Farming Simulator steps into the metaverse
Why is it important?
Sam Watson Jones This is absolutely fascinating and an incredibly exciting vision for the future of farming. Simulating the farm and honing farm skills in a virtual environment is one interesting aspect of this, but the real transformation will be in our ability to predict the impact of our actions. As a result of gathering ever more granular historical data, soon it will be possible to “scroll forward” and predict the potential impact of our decisions. Farming will become all about asking the right questions. Imagine it’s day 150 for a wheat crop today, if I apply this fertiliser in this amount, what could that look like in terms of a yield response? What would happen if I applied half of the fertiliser? Or I applied the same amount but in three doses? What will happen if it doesn’t rain for the next 50 days? And so on and so on. The virtual or simulated farm will become a huge part of how we take decisions on farms.
What is it?
Farmers may only have 40 growing seasons to improve their crops, but in the metaverse that opens up a wealth of opportunity. Dubbed “a virtual world that is the digital twin of our real one”, it’s an immersive space where algorithms and AI shape your destiny, allowing you to try what-if scenarios without purchasing a single grain of seed (unless you choose to do so by virtual bitcoin). An economy set to be worth $13 trillion (£11 trillion) by 2030, tech giants Microsoft, Apple, Amazon, NVIDIA and Epic Games are all looking to morph reality and agriculture is one of the key sectors in their sights. This is so much bigger than Farming Simulator, suggests The Scoop.
Smart sprayer achieves 97% herbicide saving
Why is it important?
Sam Watson Jones Very exciting technology from an exciting company that are growing fast. Their focus on weed control and corn has put them in a strong position. The challenges that I can see with this approach are firstly, as with other “see and spray” technologies, that the farmer has to guess how much herbicide to load into the tank before heading into the field. Although they might achieve 97% savings in the field, there will still be some wasted chemical until predictions are better than they are today. Secondly, they are asking farmers who have already spent somewhere approaching $1m on a sprayer to significantly modify their machine with a new boom, which may meet some resistance. The tech looks good though, and they can certainly operate at speed - I think it would be further improved by the assistance of a scanning robot…
What is it?
Greeneye Technology is offering growers in the US corn belt a precision spraying service that gets better at identifying weeds every time you take it through the crop. Designed to retrofit to existing sprayers, the Greeneye boom has 24 cameras, 144 nozzles and 12 graphics processing units (GPU). These allow the system to spot and spray weeds in either green-on-brown (fallow) or green-on-green (in-crop), with AI technology that makes it smarter with every pass. Growers in the Midwest trialling the system achieved 97% and 88% chemical usage reduction in pre- and post-emergence spraying respectively. Currently available just for corn and soybeans, there are plans to roll it out to wheat and cotton.
Robotic weed id builds its capability
Why is it important?
Sam Watson Jones Again, we have embraced this as a significant challenge to improve our own forward operating speed. It’s really impressive to see the forward speeds that they have achieved whilst retaining a sufficient resolution of image. It shows that the perceived limitations of ground based data collection units in operating at scale will not remain issues for long.
What is it?
RoboWeedMaPS is a system being developed by a consortium led by Aarhus University in Denmark that uses image analysis for automating weed inspection. Cameras can take high-resolution images with an accuracy of four pixels per mm, at speeds up to 50 km/h in the field. With data uploaded to the cloud, the system’s AI brain has now learned to recognise 27 types of weeds. The plan is to develop a system that decides not just whether to spray, but with what herbicide and at what rate.
Virtual agronomist extends its reach
Why is it important?
Sam Watson Jones This is another indicator of the market slowly shifting towards paying for outcomes, rather than just for a piece of equipment. It’s interesting to see drone and agronomy companies shifting in this way. The seed industry is another area that is looking at this closely, with the potential for payments to shift to being performance based, rather than paying per tonne of seed. Drone legislation remains a challenge in the UK and elsewhere in the world for this to become more widely adopted.
What is it?
Crop intelligence company Taranis has extended its partnership with Drone Nerds and DJI in the United States in a bid to give more farmers and retailers access to AI technology for insightful data-driven decisions. Taranis combines data captured by satellite and planes with images acquired from a nation-wide network of Drone Service Providers (DSPs) in the US. The leaf-level, sub-mm resolution of the imagery allows actionable and constantly improving insights at scale, says the company, without the need for the farmer or retailer to fly their own drone. The tie-up allows Taranis to harness the world’s largest DJI Matrice 300 (DJI M300) drone fleet for its AI-powered precision scouting.
Review focuses on robot pickers
Why is it important?
Sam Watson Jones This is an interesting read. One of the key trends that the document points out, which is key to thinking about the future of AI and robotics, is that automation is just the first phase. Whilst much of the initial focus and investment has been around looking at the way that human beings operate and then automating that behaviour, the much bigger opportunity is using these mobile computers to detect things that human beings cannot see, such as latent disease, or that human being cannot achieve at scale, such as counting and sizing every plant. It’s encouraging to see that the industry tone has completed the transition from “will it work?” to “how is this going to be applied?”
What is it?
Defra has published the results of a review exploring how the horticulture sector can make use of innovative technologies such as packhouse automation, AI enabled robotics and autonomous guided vehicles. The review sought to understand what would be required to accelerate the development and uptake of technologies that help with tasks like the picking, packing and transporting of fruit, vegetables and flowers. A new consortium to drive adoption of proven technologies, fast-tracking of new technologies and working groups to share novel harvest practices are on the cards to alleviate acute labour shortages experienced across the sector.
