Fruits Maturity Classification for Harvesting Robot

This research proposes to create a machine vision using YOLOv5 model to detect maturity of the tomatoes on farm field which categorized by raw, half-ripe, and ripe.The use of multiple color space RGB, HSV, and LUV dataset aims to overcome external disturbances such as light and shadows.The results are the machine vision can differentiate fruits by its category with minor bias and obtaining a good mAP50 rates with the greatest result is LUV color spaces model.This machine vision with custom datasets in multiple color spaces can be implemented on farm field which is considered based on the purposes, environment and external disturbances.These models are developed for autonomous harvesting robot on farm field.