Weeds in the field are the enemy for any crop as they compete with the soil nutrients and can become a host for germs and insects threatening the crops. It is therefore essential to be combated effectively and with as little financial burden as possible for the producer and as little pollution for the environment as well.
Precision agriculture paves the way through variable rate technologies so that herbicide is applied selectively in the field and not widely. To distinguish between weeds and crops in order to spray, we need the help of computer vision through the image annotation.
Weeds due to their super-dense growth, as can be shown in the above image, are difficult to be controlled by herbicides, because they are almost impossible to be identified and distinguished one by one. Image annotation overcomes this obstacle and also indicates the number of existing weeds at the same time in a field – information that is extremely interesting! Good quality annotation ensures accurate spraying through the cooperation of the software and the electronics of the sprayer and the tractor. This is crucial, as on the one hand effective weed control is carried out and on the other hand the spraying of the crop is avoided, thing that would otherwise be inevitable.
So, image annotation on weeds is particularly valuable in the accuracy of the spray in the specific target. We have to mention, that during annotation if a weed has not been properly or not at all labeled, as a result it will not be sprayed.
Finally, the good quality of the annotation together with the rest of the spraying tools, e.g. sensors, cameras, GPS, remote sensing, ensure that the herbicide is applied to the target weed and not next to it, where it may be the crop.