کنفرانس مهندسی برق ایران

صفحه اصلی / سی و چهارمین کنفرانس بین المللی مهندسی برق

A Contrast-Aware Real-Time Defect Detection Framework for PCBs Using CLAHE-Enhanced YOLOv8

نویسندگان :

Mehran Kushesh¹ Mohammadreza Hassan Nejad Biballan²

1- دانشگاه بین المللی امام خمینی (ره) 2- دانشگاه بین المللی امام خمینی (ره)

کلمات کلیدی :

PCB Defect Detection،Automated Optical Inspection (AOI)،Deep Learning،Real-time Inspection،YOLOv8،CLAHE

چکیده :

Abstract Automated Optical Inspection (AOI) of Printed Circuit Boards (PCBs) faces significant challenges due to the minute scale of defects and complex background patterns. Traditional methods often struggle with low-contrast micro-defects, while standard detectors frequently trade off recall for speed. In this paper, we propose a novel hybrid defect detection framework that integrates Contrast Limited Adaptive Histogram Equalization (CLAHE) with the YOLOv8-Medium architecture to enhance defect visibility under uneven illumination. Experimental results demonstrate that the proposed method achieves a Mean Average Precision (mAP@50) of 95.0%, significantly outperforming traditional VGG16-based classification and standard YOLOv8-Nano models. Furthermore, the system operates at an inference speed of 18.0 FPS, making it a highly robust and efficient solution for real-time industrial quality control.