Visual detection scheme for medical rubber parts

The rubber visual inspection system adopts advanced image visual inspection technology to achieve real-time integrity detection of seals. The image processing system preprocesses each image, relevant size measurement and other operations, and compares it with the standard template image or the relevant parameters set. If the inconsistent area is found and exceeds the set parameter range, the waste alarm information and elimination control signal will be output to ensure the quality of the sealing ring. 

 For machine vision light sources, it is impossible to find a light source that can be applied to all rubber parts. However, if only one product can be applied to a visual system, it may not be cost-effective. Therefore, the key is to consider as many applicable products as possible. Or only make those rubber parts with relatively large output. Visual inspection items on small quantities of rubber parts may not be cost-effective. 

 For industrial cameras, it is better to have a higher resolution for the detection of rubber products. With higher resolution, you can get more details and detect smaller defects. Industrial cameras with 5 megapixels are usually considered. Of course, the higher resolution may not be very helpful to the project itself, because the product itself does not have such high accuracy, so the resolution is just right. 

 There are not too many requirements for industrial lenses, and the use of ordinary CCTV lenses can basically meet the requirements. For some products with through holes, when they are relatively large, you can consider the telecentric lens. 

 Then there is the image processing algorithm, which is usually circular, so you can consider using the ring ROI tool, and then use particle analysis to deal with it. If there are defects, scars, stains, spots, etc., most of the defects can be detected. Of course, the premise is that the light source lighting can make the characteristics more obvious. If the defect feature is not much different from the background feature, the system stability may not be satisfied. Edge can generally consider curvature, caliper, edge point detector and other algorithms. It mainly depends on the specific testing items. Some requirements are relatively simple, while others are more complex. 

 Detection case 

 Detection requirements: black spots, burrs, size and other appearance defects. 

 Detection effect 

 Black spots 

 Burr