Button Cell Characterization Testing Technology Research Based on Computer Vision

Article Preview

Abstract:

The use of computer vision to detect and recognize target objects in the industrial field has high academic value. In this paper, computer vision is used to identify cell’s label and detect the surface scratch. Using inter variance threshold method of image acquisition threshold processing, to obtain a better image binarization , subsequent using characterization information to establish a template, and the template using mathematical morphology operations. Object matching is achieved by matrix change of the template , in order to achieve the detection. The experiments show that this method has higher accuracy and better real-time, so online detection can be used for battery characterization.

You might also be interested in these eBooks

Info:

Periodical:

Advanced Materials Research (Volumes 706-708)

Pages:

1856-1861

Citation:

Online since:

June 2013

Export:

Price:

Permissions CCC:

Permissions PLS:

Сopyright:

© 2013 Trans Tech Publications Ltd. All Rights Reserved

Share:

Citation:

[1] Suresh BR_ Fundakowski RA_ Levitt TS. A real-time automated visual inspection system for hot steelslabs IEEE Trans Pattern Machine Intell, 1983, 5 (6):563-572

DOI: 10.1109/tpami.1983.4767445

Google Scholar

[2] Brosnan, Tadhg, Sun Da-Wen. Inspection and grading of agricultural and food products computer vision systems-a review. Computers and Electronics in Agriculture, 2002,36:193-213

DOI: 10.1016/s0168-1699(02)00101-1

Google Scholar

[3] HanBin, liu to Ann, with WangShi. Based on image processing of the printing defect computer automatic detection. Computer applications, 2002, (3) : 37-39

Google Scholar

[4] Brosnan, Tadhg, Sun Da-Wen. Inspection and grading of agricultural and food products computer vision systems-a review. Computers and Electronics in Agriculture, 2002,36:193-213

DOI: 10.1016/s0168-1699(02)00101-1

Google Scholar

[5] Mahmoud I. Khalil, Mohamed M. Bayoumi. Invariant 2D object recognition using the wavelet modulus maxima. Pattern recognition letters,2000,21(9):863-872

DOI: 10.1016/s0167-8655(00)00046-5

Google Scholar