Defect inspection of medicine vials using LBP features and SVM classifier

摘要: During the pharmaceutical process, it is inevitable that various defects emerge in medicine vials which may greatly affect product quality and reduce productive efficiency. To address these problems, a method based on feature extraction machine learning developed for vial defect inspection. On image preprocessing, we used threshold algorithm to acquire region of interest (ROI) comprised some small patches obtained through blocking, exhibiting favorable performances compared existing segmentation methods. In following computational framework, LBP descriptors are firstly extracted ROI followed by generation visual dictionaries application k-means clustering. Since can essentially represent image, finally employ support vector (SVM) classifier inspect whether with flaws. procedure extraction, experiments show yields superior performances, (maximum recognition efficiency about 90%) others, owing exact texture features.