Automated Defect Detection in Cigarette Manufacturing Using Improved YOLOv8 Model

Cigarette defect detection is a critical process in tobacco manufacturing to ensure product quality and reduce defect rates. This paper presents a comprehensive study on automated defect detection for two popular cigarette brands: Furongwang and Changbaishan. The dataset used in this study contains six common defect types, including puncture, foam inclusion, wrinkle, misalignment, open end, and powder attachment. To address these challenges, we propose a deep learning-based approach utilizing the YOLOv8 object detection model, which is specifically fine-tuned to automatically identify these defects with high precision and efficiency. The YOLOv8 model, equipped with advanced feature extraction capabilities, has shown remarkable performance in accurately detecting multiple defect types in real-time production environments. The experimental results demonstrate that the proposed model achieves a detection accuracy of over 98%, showcasing strong generalization performance across various defect types. This study highlights the practical value of using advanced AI solutions to enhance quality assurance in cigarette manufacturing, ultimately reducing labor costs and minimizing human error.