Recovering a 4-DOF SCARA robot arm integrated with machine vision for real-time product classification

This study proposes a real-time automated sorting solution by integrating a recovered 4-DOF robot arm with a machine-vision algorithm. The inverse kinematic computation is integrated in a PLC to control the robot arm. A QR code–based image processing method is used to decode product information, localize target positions, and transmit coordinates to the PLC for the sorting process. Experimental results confirm the successful recovery of the robot arm, achieving average positioning errors of 0.24 mm and 0.12 mm on the X- and Y-axes, respectively. The integrated system effectively classified and sorted three types of QR-labeled product boxes with an accuracy of 93.33% and a processing time of 16 s per product. The proposed potential solution is entirely suitable for small and medium-sized enterprises.