Research on Printing Icon Defect Detection Based on Variation Model and Machine Learning

A printing icon defect detection method based on variation model and machine learning is proposed to address the shortcomings of defect detection methods for glass bottle printing on industrial production lines. On the basis of using the standard variation model for bottle icon detection, subjective measurement standards are introduced from two perspectives to construct a locally variable variation model, and support vector machine in machine learning is combined to address the false detection problem in the detection process. The experimental results show that compared with existing methods, the research method proposed in this paper can more accurately detect defective bottle icons, and is more subjective and efficient in practical applications.