LAUSR

Abstract The vial, a bottle known to store the drug, should be controlled to meet the requirements of the standard dimension. Due to problems with a visual inspection, there is a need to develop an automated inspection system. In this paper, a machine vision system for measuring and controlling the dimensional characteristics of medical glass vials has been developed. In this regard, because of the difficulty of taking images of glass vials and reflecting the light that may have these images, some innovative actions have been taken to determine the way for obtaining the appropriate images. Also, the effectiveness of several common segmentation methods has been examined and a heuristic segmentation method is proposed to extract vial borders. Finally, using to integrate heuristic segmentation method and appropriate post-processing methods as well as employing machine learning, an automated approach for measuring different dimensional characteristics of vials is proposed and evaluated by real samples.