A Defect Detection Algorithm for Optoelectronic Detectors Utilizing GLV-YOLO.

Area of Science:

• Materials Science
• Electrical Engineering
• Computer Vision

Background:

• Photodetectors are critical components in various applications.
• Accurate surface defect detection is essential for photodetector manufacturing and performance.
• Existing methods may lack the real-time processing and precision required for industrial settings.

Purpose of the Study:

• To develop an optimized algorithm for real-time and accurate photodetector surface defect detection.
• To improve the efficiency and performance of defect detection models in manufacturing environments.

Main Methods:

• An optimized GLV-YOLO model was developed for photodetector defect detection.
• Incorporated the GhostC3_MSF module to reduce model complexity and parameter count.
• Integrated the LSKNet_3 attention mechanism for enhanced feature extraction.
• Utilized the WIoU loss function to improve generalization performance by minimizing geometric penalties.

Main Results:

• The proposed algorithm achieved 98.9% accuracy.
• The model has 2.1 million parameters and a computational cost of 7.0 GFLOPs.
• Demonstrated superior performance and efficiency compared to existing methods.

Conclusions:

• The optimized algorithm effectively meets the demands for real-time and precise defect detection in photodetector manufacturing.
• The integration of GhostC3_MSF, LSKNet_3, and WIoU significantly enhances model performance and efficiency.
• This approach contributes to improved quality control and advancement in photodetector technology.