A hybrid vision-based surface coverage measurement method for robotic inspection. (June 2019)

• Record Type: Journal Article
• Authors: Shahid, Lubna; Janabi-Sharifi, Farrokh; Keenan, Patrick
• Is Part Of: Robotics and computer-integrated manufacturing. Volume 57(2019)
• Page Start: 138
• Page End: 145
• Publication Date: 2019-06
• Subjects: Image segmentation -- Vision-based surface coverage measurement -- Morphological operations -- Graph-cuts -- Robotic shot peening -- Cell segmentation -- Coin recognition -- Robotic inspection; Robots, Industrial -- Periodicals; Computer integrated manufacturing systems -- Periodicals; Robotics -- Periodicals; Robots industriels -- Périodiques; Productique -- Périodiques; Robotique -- Périodiques; 670.285
• DOI: 10.1016/j.rcim.2018.11.009
• Languages: English
• ISSNs: 0736-5845

Abstract

Highlights: A hybrid method based on morphological operations and Boykov graph-cuts segmentation is proposed for vision-based coverage measurement in shot peening. The results obtained showed that the proposed method outperformed the other implemented segmentation techniques (by the authors) for shot peen coverage measurement. The developed method is generic and its other applications including for blood cell and coin segmentation are demonstrated. Abstract: Vision-based surface coverage measurement serves a large area of applications including robotic shot peening, robotic inspection in the fields of blood cell segmentation, coin recognition systems and many more. Manual visual inspection is time consuming and prone to human error. The objective is to develop a real-time generic algorithm, robust to non-uniform illumination with high accuracy, and having a relatively simple experimental setup. In this paper, a hybrid method based on morphological operations and Boykov graph-cuts segmentation is proposed for vision-based surface coverage measurement. The primary focus is on shot peening; however, the proposed algorithm will be extended towards other applications such as blood cell segmentation and coin recognition. The performance of the proposed algorithm is evaluated experimentally.