A Novel Edge Detection and Localization Method of Depalletizing Robot

The application of intelligent robots to perform the depalletizing task is a common requirement in warehouse automation. To solve the problem of identification and localization caused by the disorderly stacking of boxes in pallet, and to eliminate the interference of the reflective material contained in the stacks, this paper proposes an edge extraction algorithm that combines 3D and 2D data. The algorithm firstly obtains the plane position data through three-dimensional point cloud, secondly uses an edge detection algorithm to extract edges in the two-dimensional image. Finally, an optimal segmentation strategy is performed, which is based on the results of point cloud segmentation, edge extraction, and the size information of boxes. Therefore, we can determine the position of each box in the space accurately. Compared with algorithms that only use 2D and 3D data, our method can effectively filter interference. The accuracy rate is close to 100%, which meets the requirements of industrial applications.