Capsules Online Detection System Based On Machine Vision

In some areas of industrial production, the use of machine visiontechnology can greatly improve product quality and production efficiency.The hard capsules have been widely used in the pharmaceutical industry.Capsules are firstly filled with powder, and then the patient dose. Thereforeits quality is directly related to people’s health. With the continuousimprovement of the national drug inspection standards and the increasing theamount of capsules, traditional light inspection method for inspecting thequality of capsules has been difficult to meet the demand for capsulemanufacturer. The computer visual detection technology to replacemanual light inspection has a great business prospects. As a non-destructive,non-contact means of inspection, the machine vision can improve theefficiency of capsule production and ensure the overall quality of thecapsules.Paper has conducted in-depth research based on largely querying variousarticles of visual inspection, hardware design technology, software designtechnology and image processing algorithm technology. We have developed acapsule online inspection system based on machine vision. The inspectionsystem has reached110000per hour, and substantially contained all thecapsules defects types, such as the length, spots, bubbles, cracks, scratches,nicks, flower head, foreign matter, color, etc. The main work is as follows inthis paper:(1) Paper has presented a set of visual inspection hardware architecture whichincluding external mechanical configuration, the image acquisition system, the imagetransmission system, the image inspection system, the defect capsule removal unitand other auxiliary equipment. After the initial transmission, the capsule is directlyplaced in the chain groove. The chain maintains continuous movement. The camerawill capture real-time capsule image, the image data obtained is transferred to thecomputer memory through real-time transmission system. Machine detectionsoftware will feedback operation results by inspecting the defect of capsule image. (2) Capsule inspection software based on machine vision has been developed inthis paper. The software uses Microsoft’s Visual Studio2010as developmentplatform, MFC as an interface, OpenCV2.3.1for the development of imageprocessing algorithms library. The software’s main functions include systemoperation control, inspection parameters, real-time image display, teinspection results,and other features of the feedback. Its simple and friendly interface is veryuser-friendly management.(3) This paper designs a set of capsules segmentation, inspection algorithm. Bythe initial image segmentation, capsules will then be segmented into a single capsuleimage. There are defects inspection algorithm which including the frontal capsule,capsule head and capsule color. Ensure capsules to be inspected in360degree. Paperdesigns an image positioning and color inspectiong algorithm for uniformillumination conditions.Finally, research result is summarized on the subject. It elaborates the insightsand harvest in the course of researching. Study also points out the inadequacies of thework and study to be improved in the future work.