Related Experiment Video
Updated: Sep 21, 2025
03:31
Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications
Published on: December 15, 2023
652
Tire Speckle Interference Bubble Defect Detection Based on Improved Faster RCNN-FPN.
Shihao Yang1, Dongmei Jiao1, Tongkun Wang1
1College of Mechanical and Electrical Engineering, Qingdao University of Science and Technology, Qingdao 266061, China.
Sensors (Basel, Switzerland)
|May 28, 2022
Summary
This study introduces an improved Faster RCNN-FPN network for detecting tire crown bubble defects. The enhanced model boosts detection precision for small, low-contrast defects, improving overall tire quality control.
Area of Science:
- Computer Vision
- Deep Learning
- Materials Science (Tire Industry)
Background:
- Object detection using deep learning is advancing rapidly.
- Tire defect detection faces challenges like low contrast and small defect scales.
- Existing methods struggle with precision in identifying tire crown bubble defects.
Purpose of the Study:
- To enhance the precision of detecting tire crown bubble defects.
- To address limitations in current deep learning-based object detection for tires.
- To improve the identification of small-scale, low-contrast defects.
Main Methods:
- Proposed a novel feature pyramid network (FPN) integrated with Faster R-CNN.
- Implemented feature fusion across multiple levels and directions.
- Utilized a tire crown bubble defect dataset for validation.
Main Results:
- The improved Faster RCNN-FPN demonstrated enhanced small object detection and localization.
- Achieved a 2.08% increase in mAP [0.5:0.95] and a 2.4% increase in AP0.5.
- Significantly improved the precision of detecting tire crown bubble defects.
Conclusions:
- The developed network effectively overcomes challenges in tire defect detection.
- Feature fusion across levels and directions is crucial for small object detection.
- The enhanced model offers a promising solution for automated tire quality inspection.
