One-stage defect detection method for plastic caps

This paper presents a plastic cap defect detection model. Plastic caps play a crucial role in industrial production, but they are susceptible to various defects caused by factors such as raw materials and manufacturing processes. Traditional defect detection methods rely on complex feature engineering and classifiers, leading to limited accuracy. To overcome these limitations, this study proposes a defect detection solution that leverages the YOLO model's renowned fast and end-to-end detection capability. By training on a substantial datasets of labelled plastic cap images, an efficient and accurate defect detection model is constructed. Specifically optimized for plastic cap defects, the model achieves a remarkable accuracy of 96% with low false positive and false negative rates. Comparative experiments and evaluations validate the superior efficiency and accuracy of the proposed method compared to traditional approaches. Consequently, this study presents a highly effective solution for plastic cap defect detection.