A framework for printing, detection, tracking, and registration of deformable grids. Issue 1 (2nd January 2017)

• Record Type: Journal Article
• Authors: Kenyon, Tyler S.; Spence, Allan D.; Capson, David W.
• Is Part Of: Computer-aided design and applications. Volume 14:Issue 1(2017)
• Page Start: 48
• Page End: 57
• Publication Date: 2017-01-02
• Subjects: Machine vision -- blob feature detection -- dot grid printing and tracking; Computer-aided design -- Congresses; Computer-aided design -- Periodicals; Engineering design -- Data processing -- Congresses; Engineering design -- Periodicals; 620.00420285
• DOI: 10.1080/16864360.2016.1199755
• Languages: English
• ISSNs: 1686-4360

Abstract

ABSTRACT: This paper proposes a novel framework for high resolution printing, detection, tracking, and registration of deformable grids. The undeformed dot feature grid is created on the target surface using a flat bed inkjet printer. After deformation, the surface is imaged using close-range monocular vision, and the dots are detected and localized to sub-pixel accuracy. A novel, deformation invariant grid registration topological map and associated algorithms are used to track features between subsequent video frames. By matching the map to the undeformed grid geometry, stretching or compression of the distance between the dots can be used to compute the material strains. Accuracy, repeatability, and computation time for several image based dot feature detectors are compared, and experimental laboratory results demonstrate consistent robustness of the method under practical imaging conditions. GRAPHICAL ABSTRACT: