Font Size: a A A

Defects Of The Pharmaceutical Blister Packaging Machine Vision Inspection Technology

Posted on:2008-11-29Degree:MasterType:Thesis
Country:ChinaCandidate:D Y XieFull Text:PDF
GTID:2208360215984895Subject:Computer technology
Abstract/Summary:PDF Full Text Request
The aluminum-plastic blister packaging, which is also named Press through Packaging (PTP) has been applied for packing drug, food, as well as other personal care products and received more and more favor by consumer for its convenience to use or take. The traditional quality inspection of PTP by eyes is time-consuming, tiring, inaccurate and subjective.So, an automated real-time visual inspection system has been developed in this paper, which is composed of imaging unit, illumination unit and supervise unit, so as to achieve the aim of desired product quality, to alleviate tiring of worker as possible and reduce product cost simultaneously. First, a unit for image capture and illumination has been developed which is able to satisfy the need of anti-jamming and real time. To get the test images which are gray level distributing uniformity, contrast suitability and whole smoothing, the methods such as histogram equalization, subsection linearity transform and ameliorated median filter has been adopted for preprocessing of sample image. The technique of gray level threshold segmentation has been used for separating tablets in test image from background and Canny operator has been used for extraction tablet edge. Based on commonness of geometry character of tablet defects, a fast reptile method with alterable windows for image boundary tracing has been ameliorated to recognize the possible defects in the tracing process simultaneously. Finally, the geometry characteristic parameters of tablet, such as area, perimeter and roundness, has been measured so as to make sure whether the blister packaging is good product or not by template matching.The experiments and running results indicate that the percentage of detecting the defects of missing, broken tablets and tablet fragments has exceeded 95% and ideal effect has acquired.
Keywords/Search Tags:machine vision, aluminum-plastic blister package, defect detection, feature extraction
PDF Full Text Request
Related items