• Laser & Optoelectronics Progress
  • Vol. 57, Issue 8, 081505 (2020)
Zhipeng Wu1, Danping Huang1、*, Kang Guo2, Jianping Tian1, Licheng Wu3, and Shaodong Yu1
Author Affiliations
  • 1College of Mechanical Engineering, Sichuan University of Science & Engineering, Yibin, Sichuan 644000, China;
  • 2Caihong (Hefei) LCD Glass Co., Ltd., Hefei, Anhui 230000, China
  • 3Hebei Economy Management School, Shijiazhuang, Hebei 0 50000, China
  • show less
    DOI: 10.3788/LOP57.081505 Cite this Article Set citation alerts
    Zhipeng Wu, Danping Huang, Kang Guo, Jianping Tian, Licheng Wu, Shaodong Yu. Optical Cable Pitch Detection Method Based on Machine Vision[J]. Laser & Optoelectronics Progress, 2020, 57(8): 081505 Copy Citation Text show less
    Diagram of system structure
    Fig. 1. Diagram of system structure
    Optical cable visual information. (a) Class I; (b) class II
    Fig. 2. Optical cable visual information. (a) Class I; (b) class II
    Commutation point and cable pitch
    Fig. 3. Commutation point and cable pitch
    Influence of different types of visual information. (a) Influence of class I sample cable point and cable ties; (b) influence of class II sample cable ties
    Fig. 4. Influence of different types of visual information. (a) Influence of class I sample cable point and cable ties; (b) influence of class II sample cable ties
    Binarization results of original images. (a) Class I; (b) class II
    Fig. 5. Binarization results of original images. (a) Class I; (b) class II
    Pretreatment process of class Ⅰ cable. (a) Original image; (b) thresholding; (c) denoising; (d) fitting
    Fig. 6. Pretreatment process of class Ⅰ cable. (a) Original image; (b) thresholding; (c) denoising; (d) fitting
    Pretreatment process of class Ⅱ cable. (a) Original image; (b) filtering; (c) grayscale stretching; (d) fitting
    Fig. 7. Pretreatment process of class Ⅱ cable. (a) Original image; (b) filtering; (c) grayscale stretching; (d) fitting
    Results of matching. (a) Success for matching; (b) failure for matching
    Fig. 8. Results of matching. (a) Success for matching; (b) failure for matching
    Template area division for different types of commutation points. (a) Model1; (b) model2; (c) model3
    Fig. 9. Template area division for different types of commutation points. (a) Model1; (b) model2; (c) model3
    Flow chart of automatically constructing template
    Fig. 10. Flow chart of automatically constructing template
    Best template
    Fig. 11. Best template
    Physical image of detection system
    Fig. 12. Physical image of detection system
    Cable pretreatment results. (a) Class I; (b) class II
    Fig. 13. Cable pretreatment results. (a) Class I; (b) class II
    Matching results of template partition precise positioning method
    Fig. 14. Matching results of template partition precise positioning method
    Numerical results of different matching methods and different types of pitch errors. Direct template matching for (a) class I and (c) class II; template partition precise positioning method for (b) class I and (d) class II
    Fig. 15. Numerical results of different matching methods and different types of pitch errors. Direct template matching for (a) class I and (c) class II; template partition precise positioning method for (b) class I and (d) class II
    Cable point gray
    116.05111.9497.86
    Errorrate /%AveragegrayErrorrate /%AveragegrayErrorrate /%Averagegray
    17.516718.019420.0171
    15.017715.520216.0188
    10.018711.021011.0205
    5.51973.02184.5213
    9.014.016.02072172277.013.019.022623424310.014.018.0225236250
    Table 1. Relationship between mean value of the template and matching error rate at different node gray level
    Error rate/%00100
    Matrix100×20120×24150×30200×40220×44
    Cable pointgray116.05111.94106.3397.8690.36
    Table 2. Relationship between v' of optimal filtering structural elements matrix and mean values of node gray level
    MethodCablepitch /pixelVariance /pixelErrorrate /%
    Original6887015.0
    Precise positioning71130
    Table 3. Comparison of common template matching and template partition precise positioning method
    ClassLength /mCable pitch /pixelVariance /pixel
    I10006944
    II10007113
    Table 4. Average pitch and error results measured by two types of optical cables
    Zhipeng Wu, Danping Huang, Kang Guo, Jianping Tian, Licheng Wu, Shaodong Yu. Optical Cable Pitch Detection Method Based on Machine Vision[J]. Laser & Optoelectronics Progress, 2020, 57(8): 081505
    Download Citation