Research On The Online Inspention Of Medical Tube-Tybe Bottle

There are cracks, bubbles and other defects in the glass bottle production. In recent years, the detection of glass bottle production in the country is manual or semi-manual mostly. In the meanwhile, workers endure tremendous strength, easily affected by the work environment and mood. So, the test result is unstable. To not disturb the production, online defect detection system completes the detection both the mouth and body parts, and removes the unqualified in the bottle movement.Light changes over time and much noise light disturbs our work. So lighting in the darkroom is selected. We choose image filter to reduce unexpected noise, after we get the bottle picture. To bright the dark picture, histogram normalization is necessary. Next, the binarization processing which separate the target as the outlook and the rest as the background could extract our interested part. Glass defect detection is divided into two parts. The center of the bottle mouth is tested by the four-point calibration center vertical intercept method and dual circles are used in the defect detection. In the body part, length and density defined will distinguish the strip and near-round defects. The number of pixels contained by the defective parts is counted and compared with the preset threshold. A bottle will be considered unqualified if the former is more than the later.Experimental results show that the image quality is improved after the image preprocessing. And distinguishing the defects improves the rate of identified the unqualified bottle.