Towards Sustainable and Green Agriculture: Integrating Machine Learning and Fuzzy Rough Set Analysis to Enhance Fruit Classification and Ripeness Detection

Sustainable and green agriculture aims to incorporate environmentally conscious practices into the growing and harvesting of crops, including fruit production. Nowadays, a lot of technologies are developed for agricultural applications, and the majority of them are used to classify and assess the maturity of fruits. Measuring the fruit’s maturity level is essential to obtaining fruit of the highest quality and a crucial step in guaranteeing fruit quality in the agricultural supply chain. At the moment, professionals perform this process manually, which takes time and is prone to errors. To address this, various automatic methods using machine learning and deep learning techniques have been proposed, which can analyze raw data and eliminate the need for complex engineered features. This study covers the idea of classifying fruit maturity using information on the composition of minerals and the content of a heavy metal in in three distinct fruit flours derived from unripe and ripe (peel and pulp) using a machine learning algorithm and feature selection based on a Fuzzy Rough Set method.