Modelo de detecção e identificação de doenças em plantas agrícolas baseado em machine learning: caso da Mandioca

Abstract

This study addresses the detection and identification of cassava diseases using Machine Learning techniques, contributing to food security and the sustainability of the agricultural sector in Mozambique. Agriculture plays a fundamental role in the Mozambican context, accounting for approximately 25% of the Gross Domestic Product (GDP) and employing about 70% of the national workforce, particularly smallholder farmers. However, agricultural production faces significant challenges due to the occurrence of pests and diseases, which can lead to production losses exceeding 50%, compromising farmers’ income and food security. In this context, the study proposes the use of machine learning models for the early detection of diseases in cassava leaves, enabling faster and more efficient diagnosis, as well as reducing the indiscriminate use of agrochemicals. The research includes an analysis of the main diseases affecting cassava crops, along with the development and evaluation of Machine Learning models suitable for the specific conditions of the Mozambican context. Additionally, a functional prototype was developed following an agile approach, including requirements gathering and prioritization, with the aim of providing a practical application tailored to farmers’ needs. The obtained results demonstrate the potential of artificial intelligence techniques to improve the accuracy and speed of agricultural disease diagnosis, contributing to the modernization and sustainability of agriculture in Mozambique.