Drug Blister Packaging Defects Detection Based On Machine Vision Technology

The research adopts machine vision technology to achieve pharmaceutical blister packaging defect detection problem. To build a set of machine vision hardware system, through image acquisition, image analysis and processing, image recognition, eventually detect whether there is a lack of piece in the pharmaceutical blister and pieces, and then to determine whether there is unqualified pills blister packaging. Ultimately design a set of defect detection software system, can realize the function of defect inspection of pharmaceutical blister images.First of all, the paper introduces the light source, optical lens, industrial camera and image acquisition card and other related machine vision lighting system, the functions such as image acquisition system types, specifications and other relevant information, reasonably set up a set of used for image acquisition, image analysis and processing, image recognition, and machine vision system hardware platform of software development.Then, the paper based on the image obtained by the hardware system, to adopt gray-scale mode transformation, histogram equalization operation to enhance contrast, median filtering, and morphology processing operation to complete the image preprocessing algorithm. The following is used the image after pretreatment followed by template matching based on the shape of the machine vision recognition algorithm to complete the image analysis and detection, and judge whether the image is flawed, and then determine whether the blister packaging is qualified.In the end, this paper adopts the method of Visual C++6.0and Halcon joint programming to design a set of pharmaceutical blister packaging defects detection based on machine vision software systems. By Halcon powerful professional image processing function library, rich image analysis processing algorithm and human visual programming interface to complete the analysis of the tested drug image processing and detection, and eventually export code; And then through the Visual C++6.0read Halcon export image processing code, add controls and associated Halcon image processing algorithms in function, realizes the encapsulation and compile of the software system.