Agriculture is dealing with numerous challenges of increasing production while decreasing the amount of chemicals and fertilizers used. The intensification of agricultural systems has been linked to the use of these inputs which nevertheless have negative consequences for the environment. With new technologies, and progress in precision agriculture associated with decision support systems for farmers, the objective is to optimize their use. This review focused on the progress made in utilizing machine learning and remote sensing to detect and identify crop diseases that may help farmers to (i) choose the right treatment, the most adapted to a particular disease, (ii) treat diseases at early stages of contamination, and (iii) maybe in the future treat only where it is necessary or economically profitable. The state of the art has shown significant progress in the detection and identification of disease at the leaf scale in most of the cultivated species, but less progress is done in the detection of diseases at the field scale where the environment is complex and applied only in some field crops.
Read full abstract