FP-YOLOv8: A New Algorithm for Detecting Surface Defects on Brake Pipe Ends
Author Information
Author(s): Rao Ke, Zhao Fengxia, Shi Tianyu
Primary Institution: School of Mechanical and Power Engineering, Zhengzhou University, Zhengzhou, China
Hypothesis
Can the FP-YOLOv8 algorithm improve the detection of surface defects on brake pipe ends compared to existing methods?
Conclusion
The FP-YOLOv8 algorithm significantly enhances defect detection accuracy while reducing computational costs.
Supporting Evidence
- FP-YOLOv8 achieved a mAP50 of 89.5% and an F1-score of 87%.
- The model reduced parameters and computational costs by 14.3% and 21.0%, respectively.
- Detection accuracy improved for cracks, scratches, and flash defects by 5.5%, 5.6%, and 2.3%.
Takeaway
This study created a new computer program that helps find tiny flaws on brake pipes, making it faster and more accurate.
Methodology
The study used a dataset of images to train and test the FP-YOLOv8 algorithm, which incorporates new modules for improved accuracy and efficiency.
Potential Biases
The reliance on high-quality labeled data for training may limit the model's generalizability.
Limitations
The model may struggle in complex real-world scenarios with intricate backgrounds and variable lighting conditions.
Statistical Information
P-Value
p<0.05
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website