Using MLP Neural Networks to Detect Late Blight in Brazilian Tomato Crops

Abstract

The food quality is a major issue in agriculture, economics, and public health. The tomato is one the most consumed vegetables in the world, having a significant production chain in Brazil. Its culture permeates many economic and social sectors. This paper presents a technological approach focused on enhancing the quality of tomatoes crops. The authors developed intelligent computational strategies to support early detection of diseases in Brazilian tomato crops. Their approach consorts real field experiments with inexpensive computer-aided experiments based on pattern recognition using neural networks techniques. The recognition tasks aimed at the identification foliage diseases named late blight, which is characterized by the incidence of brown spots on tomato leaves. The identification method achieved a hit rate of 94.12%, by using digital images in the visible spectrum of the leaves.