April 2021

32 n MACHINE VISION April 2021 www.drivesncontrols.com Self-learning inspection helps to set up face-mask production line rapidly U nivent Medical, a face-mask manufacturer based in Baden- Württemberg, Germany, recently established a new production line to respond to the urgent demand for FFP2 masks for medical personnel during the Covid- 19 pandemic. The line was developed with funding from the German government that helps manufacturers to adapt their production lines to produce emergency supplies. Univent decided it needed to use machine vision to inspect the quality of its masks. “Manual inspection carries a very high error rate, and we can’t allow that to happen when producing critical emergency supplies, such as face masks,”explains the company’s operations manager, Jürgen Eichinger.“Quality is at the core of all our operations, which is why we needed a flexible machine vision solution that would be quick to set up and easy to operate.” In particular, the vision system had to be able to detect defects in the masks’ultrasonic soldering, metal nose holders, strap connections, as well as the CE-mark and company logos. Inspection of the metal nose holder is particularly critical, because defective metal strips could permanently damage a cutting machine on the production line. Univent chose an autonomous machine vision (AMV) system developed by Inspekto, a German-Israeli company which has pioneered this technology and claims that it is the first self-contained, out-of-the-box vision inspection system for industrial quality assurance. AMV is a hybrid technology that merges computer vision, deep learning and real-time software optimisation technologies. Unlike traditional machine vision technologies, which are custom-built and need complex, time-consuming integration, users can install the plug-and-play AMV systems themselves on production lines without needing machine vision expertise During set-up, the user simply switches on a controller and ensures that the field of view covers the area to be inspected. They then place a good sample item in this area and use a mouse to mark a region of interest for the system to detect defects in. Inspekto argues that its technology, which requires only final integration on a production line, is ideal for manufacturers – such as Univent – which need reliable quality assurance fast, and cannot wait for the lead times of traditional machine vision projects, which can take several weeks or even months to be developed and integrated. The system can learn the characteristics of almost any new product in about an hour, from just 20 good samples. It is then ready to flag up any abnormalities during inspections. Univent’s system quickly learned the characteristics of its FFP2 masks, and was able to perform accurate and reliable quality assurance, flagging up defective masks, as well as metal strips that could damage equipment. If the company decides to switch production again in the future, the same system could be adapted quickly to inspect a completely different product. n A German face-mask manufacturer has turned to an autonomous machine vision quality inspection technology to help get a new FFP2 line up and running quickly to cope with pandemic demand. Univent was able to set up its inspection system rapidly to help spot any defects in its face-masks