Real-Time Positioning and Orienting of Pallets Based on Monocular Vision

A vision-based method for positioning and orienting of pallets is presented in this paper, which uses a single camera mounted on the fork carriage of an autonomous forklift. The orientation of a standard pallet without any fiducial marks is first determined by using a back-projection technique that projects the upper and the lower lines of the pallet front plane in input image backward onto a virtual plane with arbitrary orientation in 3D space. When the back-projected lines are parallel, the orient of the virtual plane is the same as that of the pallet front plane. The position of the pallet is then determined by using the height information of the standard pallet. Through a test on a set of real pallet images, we found that the proposed method was applicable to pallet engagement in real environment practically.