Classificação da Doença da Folha da Mandioca utilizando Redes Neurais Convolucionais

Abstract

As the second largest carbohydrate supplier in Africa, the cassava is an essential food security culture cultivated by small farmers because it can resist adverse conditions. At least 80% of family farms in Sub-Saharan Africa cultivate this starchy root, but the diseases are the main source of low production. The existing disease detection methods require that the farmers request the help of government-funded agricultural experts to visually inspect and diagnose the plants. This suffers for being too laborious, with low supply and expensive. That said, this research aims to build a predictive machine, using Deep Neural Networks (CNN) it'll be possible to identify the pathogen affected in the Cassava leaves. To guide our research the SEMMA methodology was used. The results obtained with the built algorithm were satisfactory, reaching the result of 91% of assertiveness in diagnosing pathogens.