Automatic Monitoring of Fruit Ripening Rooms by UHF RFID Sensor Network and Machine Learning

TL;DR

This paper presents a non-destructive RFID sensor network combined with machine learning to automatically monitor and classify avocado ripening stages with high accuracy, advancing precision in food quality control.

Contribution

It introduces a novel RFID-based system integrated with SVM algorithms for real-time, non-invasive ripening stage classification of avocados, enhancing control over ripening processes.

Findings

Ripening stages classified with over 85% accuracy

RFID sensors enable non-destructive monitoring

Supports precise management of ripening rooms

Abstract

Accelerated ripening through the exposure of fruits to controlled environmental conditions and gases is nowadays one of the most assessed food technologies, especially for climacteric and exotic products. However, a fine granularity control of the process and consequently of the quality of the goods is still missing, so the management of the ripening rooms is mainly based on qualitative estimations only. Following the modern paradigms of Industry 4.0, this contribution proposes a non-destructive RFID-based system for the automatic evaluation of the live ripening of avocados. The system, coupled with a properly trained automatic classification algorithm based on Support Vector Machines (SVMs), can discriminate the stage of ripening with an accuracy greater than 85%.