Research on tire crack detection using image deep learning method

Area of Science:

• Automotive Engineering
• Computer Vision
• Artificial Intelligence

Background:

• Tire quality is crucial for vehicle performance, efficiency, and safety.
• Drivers often overlook tire oxidation and other defects beyond tread depth and air pressure.
• Automated tire defect detection is needed to ensure safety and reduce costs.

Purpose of the Study:

• To design a deep learning-based method for detecting tire defects.
• To improve the traditional ShuffleNet model for enhanced tire image analysis.
• To evaluate the effectiveness of the proposed method against existing models.

Main Methods:

• An improved ShuffleNet model was developed for tire image defect detection.
• The proposed method was compared against GoogLeNet, traditional ShuffleNet, VGGNet, and ResNet.
• Tire defect detection was validated using a dedicated tire image database.

Main Results:

• The improved ShuffleNet achieved a 94.7% detection rate for tire debris defects.
• The method demonstrated robust and effective detection of various tire defects.
• The proposed approach significantly reduces tire defect detection time.

Conclusions:

• The improved ShuffleNet model is a robust and effective solution for tire defect detection.
• This technology can help drivers and tire manufacturers save labor costs.
• Enhanced tire defect detection contributes to improved vehicle safety and tire longevity.