ISBN: 978-981-11-3671-9 DOI: 10.18178/wcse.2017.06.051
Detection of Glass Fragment Based on Background Subtraction
Abstract— An authoritative quality inspection system is implemented to safety glass in China,
fragmentation test is one of the test methods. Presently, the work is finished by labor, which wastes time and
energy, it’s a reliable method to recognize crack lines of glass fragment by the means of digital image
processing of which the image method which applies background subtraction makes some improvements on
the account, in order to avoid appearing internal cavity of crack lines resulted by deduction operation, take
advantage of improved Niblack binarization method and mathematical morphology operation to refill crack
lines, then refine crack line and conduct a statistics of fragment amount. The experimental result shows that,
the algorithm not only can make a precise statistics of fragment amount, but improve operating efficiency
Index Terms— image processing, background subtraction, binarization, mathematical morphology.
School of Software, Shenyang University of Technology, CHINA
JiSchool of Information Science and Engineering, Shenyang University of Technology, CHINA
Cite: Jiang yan, Fu Jiangwei, "Detection of Glass Fragment Based on Background Subtraction," Proceedings of 2017 the 7th International Workshop on Computer Science and Engineering, pp. 294-299, Beijing, 25-27 June, 2017.