Study on On-line Quality Inspection System of Aluminum-plastic Foamed Mask Medicinal Package

摘要

This paper proposes and designs an on-line quality inspection system of aluminum-plastic foamed medicinal package based on computer vision technique. Firstly, this paper gives a brief introduction to the structure and working principle of medical package inspection system. Secondly, computer vision and automatic control techniques are adopted to build the hardware system of the quality inspection system. Thirdly, the main image processing algorithms for inspection of aluminum-plastic foamed mask medicinal package are introduced. Self-adaptive image segmentation algorithm is presented for the pretreatment of the medicinal package image. A novel region-separated method is adopted to get the features of the processed image. Therefore the features of the medicinal package image like geometric features are calculated accurately. Lastly, carries on the match algorithm between image to be inspected and the standard template image. The method is applied to one hundred of aluminum-plastic foamed mask medicinal package images with different quality conditions. The actual experiment results showed that this method can accurately inspect hollow foams, omission assemblage and so on closed with quality questions in the package of tablets. The detection accuracy is about 100%, so the method is proved to be valid.