Drone technology is improving all the time, and as uptake increases the price comes down – so what are the latest developments to hit the market?
According to Jonathan Gill, mechatronics researcher at Harper Adams University, agricultural drones are becoming increasingly accessible to owner operators, with an entry level drone and camera costing around £5,000.
“Originally you’d have had to buy the software but with high speed internet there are a good number of providers with server processing: You just upload the data and they process it.”
Currently, the main limitation to drone use is the requirement for operators to keep them within 500 metres and in direct sight.
“But manufacturers are experimenting with adding transponder units to the drones so they can be identified by other aircraft and air traffic control. That will be a lot safer, as it’s difficult for pilots to see drones; this way they will be far easier to identify.”
Technology is also being developed so drones can avoid potential collisions automatically.
“This is the first step towards being able to fly drones out of sight.” The past two years have seen a lot of data captured from drones, and companies are using that to improve the identification of pests and diseases through hyperspectral wavelengths.
“In addition, there are sniffing sensors to detect fungi in the air and gas sensors to pick up other outputs from the plants.”
Will Wells, chief executive at Hummingbird Technologies, says integrating drone and mapping technology with intelligent machine-based algorithms offers pinpoint disease and weed identification, enabling farmers of the future to make better and faster decisions.
The firm is using machine learning algorithms to tackle problems before they even appear visually – septoria detection is now 93% accurate and black-grass mapping is accurate to a high threshold, says Mr Wells.
“This is about absorbing data from images and interpreting it through farm management software to a decision-making capability.
“A drone is only as good as the map it produces. Machine learning-based software is the future.
Taught in the right way, these algorithms can unlock what 50 years of agronomy experience looks like.”