DETECÇÃO DE MOFO-BRANCO COM IMAGENS AÉREAS

Abstract

The use of remote sensing using drones has the potential to optimize the monitoring and identification of fungal diseases in agricultural crops of economic importance. In this sense, the objective was to use a low-cost, easily acquired sensor embedded in a drone to identify Sclerotinia sclerotiorum in soybeans. A flight was performed when the pods were fully developed (R4) and another one close to harvest (R8). Disease incidence and severity were georeferenced and subjected to linear regression analysis against plant reflectance in visible bands. In the first flight, carried out in the flowering stage, due to the recent application of the fungicide, it was not possible to identify what the changes were due to using the RGB sensor. In the second flight a positive correlation was obtained between the blue band and the average incidence (r²=0.34), allowing to detect alterations in the plants infected by S. sclerotiorum using the blue band of the RGB sensor, even with the limitations for determination of the previous infection, it is hoped that it will be possible with an infrared sensor and thus, implement an early management and save the culture, which was not possible in this case, since the severity only became perceptible on the second flight with an advanced condition of the pathology.