Coffee Leaf Rust Detection Using Genetic Algorithm

Abstract

In Brazil, most of the productive coffee plants is susceptible to rust, a severe disease caused by a pathogenic fungi which attacks the leaves of coffee plants, thus causing a drop in coffee production of up to 45%. To address this problem this paper proposes a genetic algorithm-based solution to identify rust in coffee leaves, thus contributing to a better combat of its fungus and less use of pesticides. We use the genetic algorithm to compute an optimal convolutional kernel mask that emphasizing color and texture features of the fungus infection in the leaf. Comparison with data provided by experts indicated that our approach represents and feasible solution for the problem of identifying rust.