We are developing an embedded camera network in our lab that mounts on agricultural sprayers. The system allows the recognition and treatment of problematic areas in real time and in precise manner. It also allows our farmers to visualize their crops on a weed map, so they can track the evolution of problems.
Our system relies on 4 main features:
1) Embedded camera network – A camera network is embedded directly on the booms of the sprayer. The cameras are compatible with all sprayers on the market, regardless of size or brand.
2) Weed recognition – During weeding, an embedded computer, linked to the cameras, analyzes the images of the field in real time. It is capable of recognizing weed species present within crops.
3) Real-time spot spraying – The on-board computer controls the spraying in real time, in order to identify and only spray on the weeds. It can cut a section of nozzles, or work one nozzle at a time.
4) Weed mapping – Information collected in the plots is automatically and immediately transcribed in the form of maps. These maps can be integrated into the sprayer for further processing or in agricultural processes management software.