Research on Automatic Cigarette Appearance Defect Detection System Based on Deep Learning Algorithm

In this paper, an automatic cigarette appearance defect detection system based on deep learning algorithm is proposed, aiming to improve the defect detection efficiency and accuracy in the cigarette production process. By using a pre-trained ResNet18 network model, the system automatically identifies defective products in the cigarette production line. We trained and validated the model on the dataset, and showed the trends of the loss function and accuracy during the training process. The experimental results show that the model can effectively recognize multiple cigarette defects, especially in handling broken and deformed defects, and exhibits high accuracy and robustness. The study shows that the application of deep learning-based defect detection technology in cigarette production has significant potential and can provide intelligent solutions for quality control.