Tobacco Leaf Grading Algorithm Based on Expert Knowledge & Deep Learning

Tobacco leaf grading is an essential process in ensuring product quality in the tobacco industry. Following the 42-grade classification system used in China, accurate grading directly impacts pricing and production efficiency. Manual grading methods are limited by their subjectivity and inefficiency. To address these challenges, this study presents a hybrid tobacco leaf grading algorithm integrating expert-defined features with deep learning models. Through extensive experiments and data visualization, the proposed method achieved a 93.2% grading accuracy, demonstrating significant improvements over traditional methods and providing a robust, interpretable solution for automated grading systems.