Czech startup Skymaps has launched “Zoneye”, an AI mannequin that detects crops and weeds from drone photos, trains on native species, and helps farmers lower enter prices by 50 % whereas boosting yields 20 %. Crop counting and germination detection additionally included – system to be offered at Agritechnica 2025.
Czech startup Skymaps has right now launched Zoneye, an AI mannequin that permits farmers to coach the system to acknowledge weeds distinctive to their area. Educated to acknowledge 37 frequent weed species, Zoneye pinpoints their actual location within the discipline utilizing drone photos. Built-in into the corporate’s CultiWise prescription-map platform, the instrument allows growers to chop enter prices by as much as 50 % whereas boosting yields by as much as 20 %.
Educated on thousands and thousands of drone-captured photos of crops and weeds, Zoneye can establish, inside minutes, all frequent weeds similar to thistle, mayweed and ragweed amongst main crops together with corn, winter wheat, soybean, sugar beet, sunflower, rapeseed, potatoes, and onions. Farmers add their very own drone photos to the cloud-based software program, which delivers ends in minutes.
Zoneye is offered worldwide to CultiWise subscribers and prices 5-20 euros per hectare. Non-compulsory CultiWise drones begin at 4,200 euros.
“Farmers have repeatedly burdened the significance of figuring out exactly the place the weeds are on the plant degree to allow them to sort out these annoying discipline invaders extra successfully whereas utilizing fewer inputs and lowering environmental affect,” stated Kornel Cziria, Chief Know-how Officer at Skymaps.
“After greater than 500,000 man-hours of growth and testing, we’re bringing Zoneye to their fields. This can be very satisfying to attain this centimeter-level accuracy,” provides Cziria.
How Zoneye works in follow
Farmers seize discipline photos utilizing drones with RGB cameras —flying at heights of 40 to 120 meters, relying on weed measurement—and add them to the CultiWise platform. The system processes photos by means of a proprietary AI mannequin educated on thousands and thousands of drone pictures.
Inside minutes, CultiWise generates exact prescription maps detailing the situation, density, and species of weeds, enabling the farmer to determine the precise quantity of herbicide required for every space. These maps are exported on to equipment terminals, guiding sprayers to use herbicides solely to weed-infested areas.
“In areas with uncommon situations, like Australia with its crimson soils, farmers can add pattern drone photos of drawback areas in actual time to CultiWise. The system rapidly adapts to the native situations, permitting what we name a ‘field-specific adjustment,’” says Cziria.
“Native situations fluctuate, each farmer’s discipline is totally different. Now we’ve got put the ability within the farmers’ fingers to allow them to prepare the system for his or her particular necessities.”
Turning maps into actions
Referred to as spot-spraying and variable price utility, this exact crop safety technique depends on prescription maps to precisely distribute herbicide. Gear from firms like Agrifac, Amazone, Horsch, and John Deere makes use of these maps to focus on crops successfully.
Many farmers already personal sprayers able to spot utility however usually lack the instruments and experience to detect weeds and generate prescription maps.
The maps inform farmers of anticipated quantity financial savings. “Zoneye gives as much as 50 % lowered herbicide enter, as a result of farmers apply solely what’s wanted and solely to the person weed vegetation. This may end up in elevated yields of as much as 20 % resulting from lowered crop stress,” provides Cziria.
Not like standard “green-on-brown” programs that detect weeds solely on naked soil, Zoneye distinguishes crops from weeds even in “green-on-green” conditions the place each seem just like the attention and to cameras.
As a result of Zoneye is so exact, it may well see proper right down to the plant degree and depend each particular person crop in a discipline. That is particularly helpful when farmers have to determine whether or not to proceed the crop or reseed lacking vegetation.
“One farm anticipated 100,000 sugar beet seeds per hectare to germinate, however many died over winter,” explains Cziria. Zoneye confirmed the zones the place over 65,000 vegetation per hectare had survived—above the financial threshold—so the farmer was capable of reseed solely the zones beneath the edge.
Utilizing the identical prescription-map know-how, Zoneye may also information seeding. By analyzing plant density and gaps within the discipline, the system can suggest adjusting seeding charges for areas the place germination is low. This ensures uniform crop institution, optimizes enter use, and helps farmers maximize productiveness throughout the sector.
Moreover, Zoneye gives early harvest estimates, giving farmers a bonus when planning assets, managing grain storage, and scheduling gross sales.
“We’re happy with Zoneye’s efficiency. It was an enormous technical problem for the software program to decipher from the drone photos what was crop and what was weed and that’s as a result of we’re coping with multitude shades of inexperienced and in that jungle, the software program should establish the form of a selected weed.”
Zoneye shall be offered at Agritechnica 2025, Nov. 9-15 in Germany on the CultiWise stand H35 in corridor 9.
