Based on machine vision technology,a PET bottle blank detection system composed of detection component,control component,conveying mechanism component and eliminating mechanism component was designed to improve the high intensity and low efficiency of manual detection of PET bottle blank defects.Halcon software was used to extract the detection area of the collected PET billet images by means of median filter denoising and adaptive threshold,and the improved Canny algorithm was applied for defect detection.The test results show that the PET bottle blank detection system can complete the detection of PET bottle surface defects such as black spots and scratches,with the test detection accuracy as high as 97.4%and the cabability of stably identifying unqualified bottles.
维普
万方数据