AUTOMATED TOMATO RIPENESS MEASUREMENT SYSTEM USING THE CNN METHOD AND LOAD CELL SENSORS FOR WEIGHT MEASUREMENT

Abstract

Tomato (Solanum lycopersicum L.) is one of the horticultural commodities with high economic value and wide consumption in society. The quality of tomatoes is determined by their ripeness level and size, which are generally assessed manually. Manual assessment requires considerable time and labor and is prone to human error due to subjectivity. This study aims to develop an automated system capable of classifying tomato ripeness levels and sorting tomato sizes efficiently and accurately. The system employs a Convolutional Neural Network (CNN) method to classify tomato images into three categories: ripe, half-ripe, and unripe, with image acquisition performed using a RaspiCam. The tomato size is determined using a Loadcell sensor, while data processing is carried out using a Raspberry Pi 5 as the main controller and an ESP32 as the supporting microcontroller. Based on testing with 120 sample data, the tomato ripeness classification achieved an accuracy of 90%, while the weight measurement using the Loadcell achieved an accuracy of 98.03%. The results indicate that the developed system can classify and sort tomatoes automatically, quickly, and reliably, making it potentially applicable for efficient post-harvest sorting in agricultural production.