Scholarlink

<p>In industrial inspection lines, precise localization of brake discs is a crucial step to ensure efficient defect detection. This study proposes an improved multi-scale template matching algorithm based on the Gaussian pyramid and similarity termination strategy, aimed at enhancing the speed of accurate localization of brake discs in industrial inspection lines to meet high efficiency requirements. After localization, the detection area is extracted using a circle fitting algorithm based on the Canny operator, and preprocessed, including median filtering and Laplace operators. Subsequently, OTSU segmentation and morphological closing operations are used to fill defects and further locate defects. For each defect pit, an accurate circle fitting is performed, and the radius of the fitted circle is converted to the actual size through pixel equivalent calibration. Finally, pit defects are screened according to the acceptance criteria. The experimental results demonstrate that the average localization time of the improved multi-scale template matching algorithm is only 8.3&#xa0;ms, which is 200.65&#xa0;ms less than the localization time of the original algorithm. The accuracy of the defect detection algorithm in this study reaches 98.5%, with a false negative rate and false positive rate of only 1.79% and 1.38%, respectively, providing more accurate and efficient theoretical support for the study of brake disc pit defects.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Brake disc defect detection method based on an improved multi-scale template matching algorithm

  • 期刊论文

  • 发表日期：

    2025年4月15日

查看全文
  • Yang Yang,
  • Jun Guo,
  • Cheng Liu,
  • Hui Qian,
  • Pinglin Gu

摘要

In industrial inspection lines, precise localization of brake discs is a crucial step to ensure efficient defect detection. This study proposes an improved multi-scale template matching algorithm based on the Gaussian pyramid and similarity termination strategy, aimed at enhancing the speed of accurate localization of brake discs in industrial inspection lines to meet high efficiency requirements. After localization, the detection area is extracted using a circle fitting algorithm based on the Canny operator, and preprocessed, including median filtering and Laplace operators. Subsequently, OTSU segmentation and morphological closing operations are used to fill defects and further locate defects. For each defect pit, an accurate circle fitting is performed, and the radius of the fitted circle is converted to the actual size through pixel equivalent calibration. Finally, pit defects are screened according to the acceptance criteria. The experimental results demonstrate that the average localization time of the improved multi-scale template matching algorithm is only 8.3 ms, which is 200.65 ms less than the localization time of the original algorithm. The accuracy of the defect detection algorithm in this study reaches 98.5%, with a false negative rate and false positive rate of only 1.79% and 1.38%, respectively, providing more accurate and efficient theoretical support for the study of brake disc pit defects.