YOLO Network For Defect Detection In Optical lenses
Habib Yaseen
TL;DR
The paper tackles the need for real-time, scalable defect detection in mass-produced optical lenses by leveraging a lightweight YOLOv8n-based system trained on a custom lens image dataset. It details data acquisition, annotation, preprocessing, and augmentation, and provides architecture-specific insights into backbone, neck, and head design tailored for edge inference. Experimental results show strong lens detection but limited defect localization, largely attributed to annotation quality and dataset imbalance, while achieving real-time performance on constrained hardware like the Jetson Nano. The work demonstrates the viability of automated optical inspection for manufacturing quality control and outlines directions for improving defect labeling and model robustness for industrial deployment.
Abstract
Mass-produced optical lenses often exhibit defects that alter their scattering properties and compromise quality standards. Manual inspection is usually adopted to detect defects, but it is not recommended due to low accuracy, high error rate and limited scalability. To address these challenges, this study presents an automated defect detection system based on the YOLOv8 deep learning model. A custom dataset of optical lenses, annotated with defect and lens regions, was created to train the model. Experimental results obtained in this study reveal that the system can be used to efficiently and accurately detect defects in optical lenses. The proposed system can be utilized in real-time industrial environments to enhance quality control processes by enabling reliable and scalable defect detection in optical lens manufacturing.