What is adaptive surface inspection?
Adaptive inspection is the process a vision system can use to update it's segmentation parameters automatically. This is the key to inspecting substrates that have a natural variation in appearance but there is still the requirement to detect low contrast defects.
Often when plastic is extruded , the appearance of the material is not uniform. This is especially evident when the substrate is back lit for a vision system. So you may ask why is this an issue? Well often this mottle effect cannot be removed , but you want your optical vision equipment to identify subtle variations or small defects. This is where self adaptive inspection is a winner.
Self adaptive inspection is where the algorithm continues to monitor the background variations. If there has not been a blemish or defect, the algorithm will update it's reference. This is often used in conjunction with an advanced secondary algorithm such as block statistics.
Self adaptive algorithms also learn the profile of the LED lighting. Any differences detected can then be fed to a classifier to check whether it is considered a defect. The image below shows a non-woven that when back lit has a varying density. The clump is highlighted by the adaptive surface inspection (ASI).
What does this enable?
Surface inspection is a must have for all roll to roll and sheet operations. It is no longer acceptable to check for gauge only. Converters need to know the severity and position of each defect. ASI (Adaptive surface inspection ) algorithms have made it possible to deploy simple but robust automated inspection solutions for the following applications.
- Plastic film extrusion
- Non-woven production
- Glass inspection
- Hot glue lay down for lamination applications
- Gravure coating applications
OneBoxVision are industry experts in the application of vision systems for the surface inspection of webs and sheets. Download our white paper that covers everything you need to know about plastic film inspection. If you have any questions or would like to learn more about this white paper do not hesitate to contact us.