Research on Online Detection Technology and Application of Leaf Cutting Width Based on Image Algorithm

In the context of intelligent and digital transformation in the tobacco manufacturing industry, the leaf-cutting process serves as a critical component of cigarette production. Traditional width detection methods for shredded tobacco primarily rely on manual visual inspection or simple visual sampling, which not only suffer from low efficiency but also exhibit significant human measurement errors, failing to meet modern tobacco production's stringent quality control standards. This study innovatively employs deep learning and computer vision technologies, utilizing the YOLOv5 algorithm to achieve automated learning and precise detection of flatness morphology and width in cut tobacco strands. The approach not only significantly enhances the accuracy and efficiency of width detection but also effectively reduces human interference in traditional methods. Building upon real-time measurement capabilities, we further explore reverse control technology for leaf-cutting width. Through statistical analysis of data deviations and feedback control models, the system automatically adjusts cutting machine parameters, ensuring stable width specifications and consistent product quality across the entire production process.