Research Of Product Detection System For Blood Collection Needles Based On Static Image Feature

With the development of the science and technology, the automatic extent of the industrial production is increasing day by day. But as the unpredictability of the production environment, the correct rate of the finished product can't be guaranteed all the time. According to the research on the position states feature of the blood collection needles which are automatically placed on the mold during the produce process, a real-time way based on the image features of machine vision, to detect the needles position automatically, is put forward in this paper.Machine vision mainly simulates the human vision with computer- from the image it picks up the information, which will be disposed, comprehended, and used in the final practical detection, measurement and control. By analyzing the needles position on the mold, the paper introduces a track and search algorithm of the feature position which will detect the needles position states on the mold. First we need to obtain the image, in which the needles have been put on the mold, by the digital camera. Then the image data will be processed with a process called pretreatment including image enhancement, image segmentation and mathematical morphology. After that the image turns into a nice binary image. By using the binary image we can track and search the feature position which will help us to determine the needles' state. Based on the feature position the next step is to judge the needles' position state. Finally the result will come out.Against the features of the image itself and real-time application, the paper works mainly on the image enhancement, image segmentation and mathematical morphology. Of course these jobs are some improvements or application promotions based on the algorithms worked by the predecessors. The needle feature position searching algorithm and the needle position state recognition algorithm put forward on the basis of application advance the effect and efficiency of the process. The judgment of feature position isn't based on the total image but a certain feature position, and then we use the relative positions among the needles to judge the needles in the small ranges of different places for three purposes: reducing the system's processing time, increasing the judgment accuracy and detecting the number of needles in the image automatically.Based on the algorithm and the construction of the system were tested for image processing algorithms and the overall system performance analysis of effectiveness and reliability. The running time of the algorithm is tested for the real-time inspection requirements and effectiveness; and the correct of the judgment handling verification system reliability. Through experiments, the paper verification by the feasibility study.