Early detection of weed in sugarcane using convolutional neural network

Weed infestation is an essential factor in sugarcane productivity loss. The use of remote sensing data conjunction with Artificial Intelligence (AI) techniques, can lead the cultivation to a new level terms weed control. For this purpose, algorithm based on Convolutional Neural Networks (CNN) was developed detect, quantify, and map weeds areas located state Alagoas, Brazil. Images PlanetScope satellite were subdivided, separated, trained different scenarios, classified georeferenced, producing information included. Scenario one CNN training test presented overall accuracy (0,983), it used produce final mapping forest areas, sugarcane, infestation. quantitative analysis area (ha) infested by indicated high probability negative impact productivity. It recommended that adequacy CNN’s for Remotely Piloted Aircraft (RPA) images be carried out, aiming at differentiation between species, as well its application detection culture crops