Related Experiment Video
Updated: Jul 10, 2026
12:26
Control of Cell Adhesion using Hydrogel Patterning Techniques for Applications in Traction Force Microscopy
Published on: January 29, 2022
6.3K
Glue Strips Measurement and Breakage Detection Based on YOLOv11 and Pixel Geometric Analysis
Yukai Lu1,2, Xihang Li3, Jingran Kang4
1Power Machinery & Vehicular Engineering Institute, Zhejiang University, Hangzhou 310014, China.
Sensors (Basel, Switzerland)
|December 31, 2025
Summary
This study introduces an advanced method for battery pack glue application quality control, integrating YOLOv11 deep learning with geometric analysis. It achieves high precision in glue dimension measurement and defect detection for new energy vehicles.
Area of Science:
- Materials Science and Engineering
- Computer Vision and Deep Learning
- Automotive Manufacturing Technology
Background:
- Quality control in new energy vehicle battery pack glue application is crucial for sealing, insulation, and stability.
- Existing detection methods struggle with complex industrial challenges like reflections, interference, and inconsistent glue strip orientations, limiting precision.
- Accurate measurement of glue dimensions and detection of defects like breaks are essential for battery pack reliability.
Purpose of the Study:
- To develop a robust and precise automated detection method for battery pack glue application processes.
- To overcome limitations of traditional vision methods in complex industrial environments.
- To enhance the quality control of glue application for new energy vehicle battery packs.
Main Methods:
- Integration of the YOLOv11 deep learning model for precise glue region extraction and interference blocking.
- Application of adaptive binarization and Hough transformation for glue strip orientation correction and image normalization.
- Utilizing connected component analysis and multi-line statistical strategies for high-precision pixel-level width and length measurement.
- Employing image slicing and pixel ratio analysis for reliable detection of glue breaks and wire drawing defects.
Main Results:
- Achieved average measurement errors of only 1.5% for glue strip width and 2.3% for length.
- Demonstrated a 100% accuracy rate in detecting glue breaks.
- Significantly outperformed traditional vision methods and mainstream instance segmentation models in precision and robustness.
- Ablation experiments confirmed the effectiveness and synergistic contribution of individual modules.
Conclusions:
- The proposed method offers a high-precision and robust automated solution for complex industrial glue application quality control.
- This approach significantly enhances the accuracy and reliability of glue dimension measurement and defect detection in battery pack manufacturing.
- The study provides valuable engineering insights for improving automated quality control in the rapidly growing new energy vehicle sector.
