Intelligent Inspection System for PP bag Production in Mushroom Cultivation

Abstract

There are automatic mushroom PP bag production machines on the market for mass production. However, the quality of the output of the machine depends on manual inspection. Thus, we design an automated optical detection and defective product replacement system for PP bags to improve the automation of mushroom cultivation to overcome the increasing shortage of labor. In the system, a camera was set up on the exit conveyor of the PP bag production machine to capture images of 12 PP bags of baskets. A five-layer convolutional neural network was used to identify the status of the 12 PP bag within 0.15 s. If there are PP bags with missing caps, the robotic arm removes the defective PP bags and replaces them with qualified PP bags. In order to make the CNN model applicable to mushroom farms, images of PP bags with four color caps and missing caps were collected. A CNN model was trained with 3072 images and tested with 480 images. The accuracy of the model was 99.58%.