It is necessary to inspect the encapsulating quality of PET bottles automatically in the high-speed production because the defects of PET bottles could directly bring food security problems. In this paper an integral inspection method of PET bottles encapsulating quality is designed based on machine vision. The defect inspection is realized on PET bottle caps with matching localization, edge extraction and contrast detection as well as other vision algorithms. Experiments proved that three kinds of defects can be detected with the proposed method in high-speed practical production lines with a speed of 67ms/frame, and the detection false alarm rate is below 0.35%.
展开▼