On-line Inspection System Of Reversed Fitting Defect On Pinhead Of One-off Injectors

One-off aseptic injector is a new kind of medical device imported on 1980's. It is widely used because of its no cross infection, low price and convenient operation. One-off aseptic injector includes needle lead and pinhead two parts. The pinhead may lead reversed fitting (pinhead insert into the needle lead) on the assembling process. This may lead direct influence on quality and even hazard people's life. Most medical device enterprises in China are using artificial visual detection to inspect reversed fitting. This method is difficult, inefficiency, poor reliability and it is easily cause visual fatigue. As the pinhead's smaller size and mass production, artificial visual detection can't ensure equality. This restricts the development of the medical device instruments seriously.Aiming at actuality, the paper developed an on-line inspection system of reversed fitting defect on pinhead of one-off injectors based on machine vision and image processing.According to the detection functions and the characteristics of pinhead, the system was designed based on machine vision, image processing and virtual instrument techonolgy. It mainly included general structure design hardware configuration, comparison and analysis on image processing methods and software design. The main research tasks were how to drive and control image acquisition system; motion control of stepping motor; design and manufacture illumination system; mothods of image processing and inspect reversed fitting of pinhead. The illumination system was designed based on scattering lighting and back lighting. It solved the problems of reflection on the surface and fuzzy boundary of pinheads, and then raised the accuracy of the system. DH-HV2002UC digital camera and computar M1214-MP camera lens were chosen based on actual situation of measured objects and existing conditions of laboratory. c[FP] controller and c[FP]-PWM-520 module made by NI Company were chosen to control stepping motor. They can take precise motion control and meet the pipeline requirements. In the image preprocessing, median filter was chosen to take image filter. Images after median filter had good edge maintenance and adaptability. After median filter, histogram modification was used for image enhancement. By compared and analyzed many kinds of threshold segmentation algorithm, OTSU threshold was used for image segmentation. Advanced morphology was used to take further treatment on binary image, including remove small particle, extract edge information by method of in gradient, image filling and particle analysis. And then inspect the reversed fitting by template matching. A complete vision detection system for one-off aseptic injector had been developed based on the platform of LabVIEW and the powerful image processing software IMAQ Vision.Experimental studies were taken at last in the paper. Experiments showed that, the detection accuracy of pinheadsφ1 .5,φ0.7andφ0.5are 99.9%,99.4% and 99%. It completely meets the on-line detection requirements.