A low-cost 2.5D robotic guidance system for depalletization of enameled iron-cast pots in a high-mix production line

Abstract

Robotic depalletization in high-mix manufacturing environments presents significant challenges due to variations in product geometry, irregular object positioning, and strict production cycle-time requirements. Conventional vision-guided robotic systems often rely on predefined object locations or expensive three-dimensional sensing technologies, limiting their applicability in cost-sensitive industrial environments. This paper presents the development and industrial validation of a low-cost 2.5D robotic guidance system for the automatic depalletization of enameled iron-cast pots in an existing enameling production line. The proposed system integrates a six-axis industrial robot, an industrial vision camera, a laser distance sensor, and a custom-designed end-of-arm tool developed through three iterative design stages. A novel zone-based inspection strategy is introduced, in which the pallet is divided into overlapping inspection regions, enabling reliable object localization despite positional deviations caused by manual palletizing and transportation. The laser distance sensor provides height compensation, while the optimized vacuum gripping system improves handling reliability for products of different masses and intermediate cardboard separators. Experimental validation was performed across 210 depalletization cycles under real industrial operating conditions, with 100 cycles dedicated to the final system. The proposed system achieves an average cycle time of 7.6 ± 0.4 s, and achieves the grasping/object localization success rate of 98%.