The ‘Need Farming’ concept put forward by Dutch machinery maker Agrifac dovetails neatly with such drastic requirements. The slightly confusing moniker means ‘farming according to need’ — where ‘need’ is the requirement of the individual plant rather than the whole field. That really is precision farming taken to another level.
Over the past few years Agrifac has developed the following three-pronged approach, under the Brilliant Simple tag: DynamicDosePlus (based on prescription maps), SmartDosePlus (injects chemical into the spray lines rather than mixing it in-tank) and AicPlus, which allows for real-time crop scanning and product application. It’s the last of these we’re exploring here.
The AicPlus camera system detects weeds via some nifty image processing and then provides the prompts for precision chemical application down to a resolution of 250mm. That level of control is apparently tight enough to spray individual weed patches rather than the whole field. It’s actually a commercial application of a camera/algorithm system developed by Bilberry, a French firm specialising in harnessing artificial intelligence for selective weed control. Berthoud is also exploring Bilberry’s technology, which in principle can be fitted to any brand of sprayer.
Back to AicPlus. Before the system can identify different weeds in a crop, it has to know what each one looks like. So an identifying algorithm must be coded for each crop and weed. This isn’t a particularly speedy process, so at the moment Agrifac’s set-up can deal with a relatively limited number of crop/weed
combinations, such as docks and dandelions in grassland or OSR in cereals. Other combos are on their way — for example, blackgrass in a cereal crop. In work the sprayer operator chooses which weed(s) to target via the system’s control box.