Abstract
The production of cork stoppers is currently a process in rapid evolution. Where in the past hand labor was common, today we observe increasing attempts to introduce technologies that increase the productivity of production lines. One these example are automated cork drillers that produce thousands of cork stoppers per hour. In order to harness this increase in productivity to its full potential, it must the followed by other processes in the production line upstream of the driller, as for example the feeding of the latter.This article presents the application of computer vision techniques that extract information of cork strips that move on a conveyor belt, to obtain automated feeding of a cork driller using a robotic manipulator. Image Processing extracts information regarding the strip position and orientation, and also which side of the strip is visible. Thus the strip is consistently placed in the driller in order to extract stoppers from the best quality cork. The segmentation of the cork strips is obtained by background subtraction. To estimate the strip visible surface, we apply Machine Learning techniques that enable a robust classification given a set of features extracted from the cork texture.In the experiences carried out, we were able to obtain 100% classification rate with a test dataset of more than a hundred cork strips.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have