EfficientNet-Lite 4-Based Classification System for Grading Philippine Strawberries

This study focuses on developing an automated grading system for Philippine strawberries using the EfficientNet-Lite 4 model and Raspberry Pi 4 hardware. The primary goal is to classify strawberries into predefined categories - Class I, Class II, and Extra Class - according to the Philippine National Standard (PNS). The study employed a dataset of 1836 images across the three classes to train and validate the machine learning model. Results demonstrate high accuracy rates in strawberry classification, with 100% accuracy for the Extra Class, 84% for Class I, and approximately 81.82% for Class II, leading to an overall accuracy of about 88.89%. This indicates the model's strong potential in automating the grading process and improving quality assessment in agricultural produce.