An Image-Based Hairline Crack Identification Method for Metal Parts

Chuanpeng Hao, Yan He, Yufeng Li, Xiaobo Niu, Yan Wang

Research output: Contribution to journalArticlepeer-review


Accurate detection and measurement of hairline cracks on metal parts, typically less than 1 mm wide, are important for remanufacturing. However, this task remains challenging due to the small crack size, high reflectance of the metal surface, and the lack of reliable real-size measurement methods. This article proposes an image-based identification method that addresses these challenges and enables precise detection and measurement of hairline cracks on metal surfaces. This method establishes a controllable imaging system with a blue dome light source to capture uniformly illuminated and high-contrast images. This setup overcomes the primary difficulty of high reflectance associated with detecting cracks on metal surfaces. Furthermore, the imaging system incorporates camera calibration to convert pixel size to real size accurately, offering significant practical value beyond mere crack localization and segmentation at the pixel level. A new hairline crack identification algorithm based on morphology is developed to extract pixels of the hairline cracks directly, followed by real-size calculation. The experimental results validate the proposed method’s effectiveness, achieving an accuracy of approximately 97% and 90% for millimeter- and submillimeter-level hairline cracks, respectively. Moreover, the average error and standard deviation surpass those of previous crack measurement techniques, demonstrating the proposed method’s superiority.
Original languageEnglish
Article number200114
Pages (from-to)1-14
Number of pages14
JournalIEEE Transactions on Instrumentation and Measurement
Publication statusPublished - 16 Oct 2023

Bibliographical note

Funding Information:
This work was supported in part by the National Natural Science Foundation of China under Grant 52175453; in part by the Innovation Group Science Fund of Chongqing Natural Science Foundation under Grant cstc2019jcyj-cxttX0003; and in part by the Graduate Research and Innovation Foundation of Chongqing, China, under Grant CYB21013.

Publisher Copyright:
© 1963-2012 IEEE.


  • Electrical and Electronic Engineering
  • Instrumentation


Dive into the research topics of 'An Image-Based Hairline Crack Identification Method for Metal Parts'. Together they form a unique fingerprint.

Cite this