Autonomous Gluing Shoe Sole Based on Pointcloud Data

Abstract

Adhesive gluing is an important step in leather shoe manufacturing, helping the shoe parts to stick together, especially between the sole and the upper body of the shoe. However, this task is still greatly relied on manual operation and robot teaching by playback. It is difficult that those tasks produce high productivity because the kinds of shoes to be performed are up to several thousands and the shoe itself is a non-rigid object, which is easy to be twisted during the manufacturing process. There are two kinds of shoe: flat objects that can be classified as 2D objects; however, for high heels, they should be considered as 3D objects. Therefore, the definition of the gluing trajectory is more complex for high heels soles. In order to ensure the satisfactory quality, the adhesive tip should be perpendicular to the sole at the point of the trajectory. This research presents a practical strategy for an automated gluing task in footwear industry by using a 6DOF industrial robot and RGB-D camera. By using 3D camera, when a random soling bond is positioned inside the robot working area, the 3D pointcloud boundary contour data of the shoe sole is extracted and interpolated for accurate and smooth trajectory generation then the manipulator will autonomously execute the trajectory tracing task. The efficiency of the system in terms of time, stability, and accuracy is proven through experiments.
Copyright © 2022 Praise Worthy Prize - All rights reserved.