Automatic detection of defective crankshafts by image analysis and supervised classification

  • Beatriz RemeseiroEmail author
  • Javier Tarrío-Saavedra
  • Mario Francisco-Fernández
  • Manuel G. Penedo
  • Salvador Naya
  • Ricardo Cao


A crankshaft is a mechanical component of an engine that performs a conversion of an alternative movement of a piston in a rotational motion of a shaft. It is a critical part and one of the most expensive of an engine. Defects in crankshafts may imply serious failures and, consequently, possible injuries and high costs. Therefore, the manufacture quality is of primordial importance for security and economic reasons. Nowadays, the quality control of crankshafts manufactured by forging in the automotive industry consists, among others, in inspecting them at the final process, using a magnetic particle procedure. This slow and highly stressful technique depends on operators and consumes many human resources, time, and space. This paper presents a methodology to automatically detect defective crankshafts. The proposed procedure is based on digital image analysis techniques, to extract a set of representative features from crankshaft images. Statistical techniques for supervised classification are used to classify the images into defective or not. The experimental results demonstrated the good performance of the proposed method with a classification accuracy over 99%, a 10% higher than the one obtained by manual inspection. Therefore, working time and personnel required for this task can be reduced when using this automated procedure.


Automotive industry Forged crankshaft Quality control Image analysis Supervised classification 



The authors would also like to thank CIE Galfor S.A. and Vigotec4 companies for their help in the experimental data collection.


This work has been partially supported by the Xunta de Galicia (Centro Singular de Investigación de Galicia ED431G/01). Additionally, the research of Ricardo Cao, Mario Francisco-Fernández, Salvador Naya and Javier Tarrío-Saavedra has been partially supported by MINECO grants MTM2014-52876-R and MTM2017-82724-R, and by the Xunta de Galicia (Grupos de Referencia Competitiva ED431C-2016-015); whilst the research of Manuel G. Penedo has been partially supported by grants Xunta de Galicia (Grupos de Referencia Competitiva ED431C-2016-047), all the previous grants through the ERDF. This work has been also supported by FORJACEMIC project (Research into new processes and micro-alloyed steels for hot forging of automotive crankshafts).


  1. 1.
    Becerra-Villanueva JA, Jiménez-Espadafor F, Cruz-Peragon M, Torres-García M (2011) A methodology for cracks identification in large crankshafts. Mech Syst Signal Process 25(8):3168– 3185CrossRefGoogle Scholar
  2. 2.
    Bradski G (2000) The openCV library. Dr. Dobb’s Journal of Software Tools 25:120–126Google Scholar
  3. 3.
    Chandna P, Chandra A (2009) Quality tools to reduce crankshaft forging defects: an industrial case study. Int J Ind Syst Eng 3(1):27–37Google Scholar
  4. 4.
    Chiou YC, Lin CS, Chiou BC (2008) The feature extraction and analysis of flaw detection and classification in BGA gold-plating areas. Expert Syst Appl 35(4):1771–1779CrossRefGoogle Scholar
  5. 5.
    Cook RJ (2008) Kappa. Wiley Encyclopedia of Clinical Trials pp 1?7Google Scholar
  6. 6.
    Daubechies I (1992) Ten lectures on wavelets, vol 61. SIAM, CBMS series, PhiladelphiaCrossRefGoogle Scholar
  7. 7.
    Fisher RA (1936) The use of multiple measurements in taxonomic problems. Annals of Genetics 7(2):179–188Google Scholar
  8. 8.
    Fontani D, Francini F, Longobardi G, Sansoni P (2001) Control of crankshaft finish by scattering technique. Opt Lasers Eng 35(6):387–396CrossRefGoogle Scholar
  9. 9.
    Francisco-Fernández M, Tarrío-Saavedra J, Mallik A, Naya S (2012) A comprehensive classification of wood from thermogravimetric curves. Chemom Intel Lab Syst 118:159–172CrossRefGoogle Scholar
  10. 10.
    Grote KH, Antonsson EK (2009) Springer handbook of mechanical engineering, vol 10. Springer Science+Business Media, BerlinCrossRefGoogle Scholar
  11. 11.
    Iborra A, Alvarez B, Jiménez C, Fernández-Merono JM, Fernández C, Suardíaz J (2000) Automated Visual Inspection system (AVI) for crankshaft production processes. Eur J Mech Environ Eng 45(1):29–34Google Scholar
  12. 12.
    Jensen FV (1996) An introduction to Bayesian networks, vol 210. Springer, BerlinGoogle Scholar
  13. 13.
    Jolliffe IT (1986) Principal component analysis and factor analysis. In: Principal component analysis. Springer, Berlin, pp 115–128CrossRefGoogle Scholar
  14. 14.
    Kotsiantis S, Kanellopoulos D, Pintelas P (2006) Handling imbalanced datasets: a review. GESTS International Transactions on Computer Science and Engineering 30(1):25–36Google Scholar
  15. 15.
    Kotsiantis SB, Zaharakis I, Pintelas P (2007) Supervised machine learning: a review of classification techniques. Emerging Artificial Intelligence applications in Computer Engineering 160:3–24Google Scholar
  16. 16.
    Li TS (2009) Applying wavelets transform, rough set theory and support vector machine for copper clad laminate defects classification. Expert Syst Appl 36(3):5822–5829CrossRefGoogle Scholar
  17. 17.
    Lindgren F, Geladi P, Wold S (1993) The kernel algorithm for PLS. J Chemom 7(1):45–59CrossRefGoogle Scholar
  18. 18.
    Mallat SG (1989) A theory for multiresolution signal decomposition: the wavelet representation. IEEE Trans Pattern Anal Mach Intell 11(7):674–693CrossRefGoogle Scholar
  19. 19.
    McEvily AJ (2004) Failures in inspection procedures: case studies. Eng Fail Anal 11(2):167–176CrossRefGoogle Scholar
  20. 20.
    McLaren K (1976) XIII The development of the CIE 1976 (l* a* b*) uniform colour space and colour-difference formula. Color Technol 92(9):338–341Google Scholar
  21. 21.
    Mitchell TM (1997) Machine learning. McGraw-Hill, New YorkzbMATHGoogle Scholar
  22. 22.
    Pater BV, Thakkar HR, Mehta SB (2014) Review of analysis on forging defects for quality improvement in forging industries. Journal of Emerging Technologies and Innovative Research 1(7):871–876Google Scholar
  23. 23.
    R Development Core Team: R (2018) A language and environment for statistical computing. R Foundation for Statistical Computing. Available at
  24. 24.
    Seber GAF, Lee AJ (2012) Linear regression analysis, vol 329. Wiley, New YorkGoogle Scholar
  25. 25.
    Tarrío-Saavedra J, Francisco-Fernández M, Naya S, López-Beceiro J, Gracia-Fernández C, Artiaga R (2013) Wood identification using pressure DSC data. J Chemom 27(12):475–487CrossRefGoogle Scholar
  26. 26.
    Thottungal AP, Sijo MT (2013) Controlling measures to reduce rejection rate due to forging defects. International Journal of Scientific and Research Publications 3(3):238–243Google Scholar
  27. 27.
    Vapnik V (2000) The nature of statistical learning theory, vol 2. Springer Science+Business Media, BerlinCrossRefGoogle Scholar
  28. 28.
    Wang K, Jiang W (2009) High-dimensional process monitoring and fault isolation via variable selection. J Qual Technol 41(3):247–258CrossRefGoogle Scholar
  29. 29.
    Wehrens R (2011) Chemometrics with R. Multivariate data analysis in the natural sciences and life sciences. Springer, BerlinzbMATHGoogle Scholar
  30. 30.
    Wold H (2006) Partial least squares. Encyclopedia of Statistical Sciences 9:1Google Scholar

Copyright information

© Springer-Verlag London Ltd., part of Springer Nature 2019

Authors and Affiliations

  1. 1.Research Group VARPA, CITIC, Departamento de ComputaciónUniversidade da CoruñaA CoruñaSpain
  2. 2.Department of Computer ScienceUniversidad de OviedoGijónSpain
  3. 3.Research Group MODES, CITIC and ITMATI, Departamento de MatemáticasUniversidade da CoruñaA CoruñaSpain

Personalised recommendations