Image Similarity Judgment Method for Waste Printed Circuit Boards

Area of Science:

• Materials Science
• Computer Science
• Environmental Engineering

Background:

• Waste printed circuit boards (WPCBs) contain valuable metals, necessitating efficient recovery methods.
• Current manual WPCB classification is labor-intensive, slow, and error-prone.
• Automated methods are needed to improve the efficiency and accuracy of WPCB recycling.

Purpose of the Study:

• To develop an image-based algorithm for automated waste printed circuit board similarity assessment.
• To enhance the accuracy and efficiency of WPCB classification for metal recovery.
• To reduce reliance on manual inspection in the WPCB recycling process.

Main Methods:

• Extraction of visual features from WPCB images, including hue, terminal region characteristics, structural complexity, and integrated circuits.
• Computation of similarity scores using weighted feature contributions.
• Development of classification strategies based on identified board-specific characteristics.

Main Results:

• The proposed image-based algorithm achieved 88.0% accuracy for targeted PCB types.
• The method demonstrated superior performance compared to a self-supervised contrastive learning approach.
• Key features like hue value and structural complexity were identified as highly effective for similarity evaluation.

Conclusions:

• An automated, image-driven WPCB similarity assessment algorithm can significantly improve recycling efficiency.
• The developed method offers a viable alternative to manual WPCB classification, reducing errors and operational costs.
• This approach facilitates streamlined metal recovery from WPCBs, contributing to sustainable resource management.