Research On The On-line Inspection System For Medical Tube-type Bottles

At present,in view of the medical regulation of glass defect detection is mainly by eye,this way the human cost and management cost is higher,and the detection speed and accuracy are unable to meet the needs of modern production.Based on the previous work,we further studied the detection of defects in bottle body of glass bottles.The main work of this paper is as follows:The principle of the hardware selection of the image acquisition system is studied.The types and characteristics of the light source are analyzed.The LED backlight source is selected as the illumination source.The relationship between the size of the imaging target surface and the resolution is known,and the optical lens is calculated and determined.Focal length;uses a backlit transillumination scheme.Through physical verification,we can see that the hardware device selection method summarized in this paper is feasible and practical.The paper analyzes the filtering principle of the traditional image denoising scheme and proposes an image smoothing algorithm based on regional consistency.The algorithm not only has good filtering effect on Gaussian noise and salt and pepper noise,but also can better protect the edge information of the image.Through experimental comparison,the proposed algorithm has better image denoising performance than the mean and median filters.Rewrite the connected domain labeling algorithm based on the previous software.The actual operation shows that the algorithm written in this paper is quite correct.