Determinação do grau de severidade da mancha branca no milho com a utilização da câmera de smartphone para captura de imagens em ambiente controlado

Abstract

Brazil is the world’s third-largest maize producer, third only to China and to the United States, therefore, maize is one of the main crops of great impact in the country. One of the obstacles to the high productivity of maize are diseases, which if they are not identified and combated can be responsible for crop losses. About 60% of maize crop losses are estimated to be caused by the white spot. In order to assist the farmer decision, a methodology was developed to determine the degree of severity of white spot on maize leaf, using digital image processing techniques. The methodology used to perform the work consists of the following steps: image acquisition, preprocessing, segmentation, postprocessing, edge detection and disease area calculation. It was verified that using the developed methodology it was possible to quantify the severity of white spot disease in maize leaf. It is concluded that the obtained result was satisfactory, since in the current literature the classification with the diagrammatic scale has a margin of error of 15% and with the methodology the error index was equal to 2.5%.