Using GHT-ICP2D machine vision to improve automated robotic sorting in industrial applications

Abstract Automated sorting is a critical step in industrial manufacturing, particularly when dealing with cluttered and unordered mass-produced workpieces. Machine vision-based sorting system demonstrated substantial processing capability and economical effectiveness; thus, it is widely adapted for automated sorting in manufacturing industries. However, machine vision systems frequently face challenges when sorting high-tolerance, high geometry complexity metallic workpieces. Sorting efficiency and accuracy can be hindered by workpiece inconsistency, environmental variations, and interference factors. To address these challenges, this study developed a novel GHT-ICP2D machine vision algorithm to enhance efficiency, precision, and robustness of the automated robotic sorting process. The algorithm was also experimentally tested with multiple types of metallic workpieces under various interference scenarios that are commonly presented in industrial sorting operations. The experimental results demonstrated that the GHT-ICP2D algorithm significantly outperforms existing methods in terms of accuracy, robustness, and computational efficiency as shown in the graphic abstract (Fig. 1). The GHT-ICP2D is a promising solution for machine vision-based automatic robotic sorting of asymmetrical, high-tolerance metallic workpieces in the manufacturing environment. For rapid industry adaptation, a user manual of GHT-ICP2D is also provided in the supplemental material.