ARES: An Automated Rotten Egg Sorter Utilizing the Egg's Physical Properties and Artificial Neural Network

Currently, the market provides egg-quality sorters that make use of spectroscopy, a technique typically employed in industrial settings. There is limited research specifically addressing the identification of rotten eggs based on weight; the researchers proposed a solution, an automated egg sorter utilizing the egg's physical properties and artificial intelligence. In order to develop a technology that caters to the needs for food sustainability, the researchers formulated a system involving two (2) subsystems that work dependently and applied together, which created a device capable of classifying and automating the sorting process of market eggs depending on its quality (good or bad). In order to develop the system, 2 subsystems were made. The first subsystem is responsible for collecting the dataset involving the egg's weight and dimensions using a load cell, raspberry pi camera, and raspberry pi 4. The second subsystem involves integrating the developed model into the raspberry pi with a gripper as an actuator. Using 3 ML algorithms, the best result, using ANN yielded an accuracy of 93.75%.