Enhancing Productivity Through Object Detection Systems on Production Conveyors in the Carton Milk

Abstract

This research investigates the application of an object detection system on a conveyor belt for sorting carton milk products. The study involves experimentation, data collection on machine-based sorting processes, and the calculation of energy consumption. The implemented system consists of a 5-meter-long conveyor integrated with industrial cameras and sensors functioning as input devices. When a product moves along the conveyor, the camera and sensor detect the object and send a signal to the PLC (Programmable Logic Controller) to verify its accuracy. The system then sends output signals to direct the carton milk products to the packaging department. The system was programmed to detect three types or colors of milk cartons and was applied to the production line in a dairy factory. The performance of the conveyor system was evaluated at three different speeds: 0.2, 0.4, and 0.6 meters per second. The detection and sorting capacities were found to be 1,200, 2,400, and 3,600 cartons per hour, respectively. The system was applied to a dairy industry case study located in an industrial estate in Phra Nakhon Si Ayutthaya Province. Based on feedback from the case study company, increasing the conveyor speed beyond 0.6 meters per second could further enhance sorting productivity and reduce production costs. The company aims to achieve a target sorting rate of 6,000 cartons per hour.