Application of Deep Learning for Mushrooms Cultivation

Abstract

In this study, we designed and implemented an intelligent system that uses image recognition to determine the status of mushrooms in real time and record the growth process. This system may help improve production efficiency and product quality. It primarily applies YOLOv4 to identify mushroom maturity and the presence of dirt; it also includes an image visual distance recognition algorithm that learns the position of the monitoring lens relative to the mushrooms. The results of the experiment demonstrate that the proposed system can accurately recognize mushroom status and calculate the distance between the lens and mushrooms.