Design and implementation of a machine learning-integrated reliable quality control evaluation system for cigarette manufacturing process

Abstract The growing needs for quality tobacco products require improvements in quality control in the manufacturing of cigarettes. Standard quality control techniques, including manual inspections and traditional automated means, may not be able to identify minute defects in the production process effectively, causing inefficiencies and inconsistency in the final product. This work suggests a machine learning-based framework that employs U-Net for image segmentation and Mountain Gazelle Optimizer (MGO) for optimization in cigarette production quality control. The proposed framework aims to enhance defect detection, decide optimally, and facilitate real-time adaptability to changing production conditions. In this paper, the Tobacco Leaf Disease Detection Dataset was utilized, which has high-resolution tobacco leaf images classified based on disease classes, covering a range of conditions from tobacco curing. The proposed method outperforms existing methods by a wide margin, with performance metrics indicating 98.12 percentage of accuracy, precision at 98.48 percentage recall at 97.74 percentage, and an F1-score at 98.11 percentage. These findings reflect the high efficiency of the framework in detecting slight defects, promoting consistency of the overall product and operational efficiency. The system also reflects the capacity to dynamically adapt to changes in the production environment for better quality and lower operational costs. Future research will investigate additional optimization methods and additional data sources for integration to enhance scalability and adaptability, further making manufacturing a sustainable and efficient process. The proposed framework is designed for real-time deployment and seamless integration with industrial quality control systems, enabling automated defect detection and adaptive process optimization. This system-level integration enhances production consistency, reduces material waste, and improves overall quality assurance in cigarette manufacturing environments.