As part of the Blue Deal, the agriculture sector must improve its handling of water. In agriculture, this means, on the one hand, better securing precipitation water and, on the other, using the available water as efficiently as possible.
However, the increase in the presence of weeds in main crops, among other things, creates competition for water. Moreover, this increase will not level off; on the contrary, the European Green Deal sets ambitious targets to reduce the use of plant protection products, among other things. This will involve alternative weed control techniques on the one hand, but also precision agriculture on the other. With the latter, it is possible, for example, to apply crop protection products only at the location of the weeds instead of the classic field application. In addition, optimisation of plant protection product efficiency is important. For this, the farmer has to know very well which weeds are present in his field.
An inventory of weeds with drones offers many opportunities in this area. Therefore, this project will put its focus on developing a model for inventorying weeds in different crops. By submitting drone images to recognition software, it should be possible to map plots on a large scale. This requires training recognition software and displaying the results in an integrated way (e.g. via Watch-it Grow). In this project, we focus specifically on the case weed tuber cyperus because of its problematic occurrence. This may be a long-term precursor for many other weeds.
