Der Blätterkatalog benötigt Javascript.
Bitte aktivieren Sie Javascript in Ihren Browser-Einstellungen.
The Blätterkatalog requires Javascript.
Please activate Javascript in your browser settings.
productronica 2021 Assembly Inspection Anzeige Intensität von LEDs Displays Armaturen-Beleuchtungen und anderen Selbstleuchtern konzipiert sind Zu den Exponaten zählt unter anderem der LED-Analyzer eFLAT-II der erste seiner Art der in der Lage ist gleichzeitig die LED-Farbe zu messen und deren Einschaltzeit millisekundengenau auszugeben Somit ist beispielweise die Qualitätsprüfung sogenannter dynamischer Blinker automatisiert im Fertigungsprozess schnell durchzuführen Das kompakte Multikanal-LED-Testsystem eFLAT-III komplettiert in diesem Jahr Premosys‘ Portfolio der LED-Analyzer Hiermit ist es möglich mehrere Module über Daisy Chain auf bis zu 200 Kanäle zu erweitern Weiterhin können LEDs vom sichtbaren Bereich bis in den Nahinfrarotbereich automatisiert vermessen werden Verdrehsichere Glaslichtwellenleiter garantieren zusätzlich dass die Lichtleiter selbst nach dem Lösen vom Messkanal immer die gleiche Orientierung aufweisen eine neue Kalibrierung und Konfiguration des Systems entfällt nw ■ 30 YEARS FUJI EUROPE CORPORATION Future innovation with NXTR-S Akey for higher productivity Success of our modular system with > 100 000 sold machines worldwide ◆ ◆ Visit us at prod uctronica 2021 Booth A3 317 FUJI EUROPE CORPORATION GmbH +49 0 6107 6842-0 | fec info@fujieuro de | www fujieuro de Machines will be instructed verbally in the future while 3D glasses offer interesting options such as instructing people during the repair a machine In addition the first applications for AI are starting to emerge in the field of image recognition The Kitov One designed for fully automated visual inspection in endofline inspection also employs AI It can for example compare a screw under inspection with a huge material library of more than 10 000 different screw types and at the same time take specific specifications into account and learn from error patterns Furthermore because the device can also independently take over and save programme optimizations made by a human it develops a certain intelligence Conversely other machines carry out defined processes on the basis of algorithms These machines record tolerances in order to evaluate them on the basis of predefined measured values However pseudo errors must be assessed by humans Some development steps are still required here whereby pure software adaptations should not be considered as AI Unfortunately the term AI is all too readily exploited for marketing purposes customers should inform themselves in advance and ask for details AI is the future but it is still a long way from becoming mainstream Yet you describe the Kitov One as a genuine AI solution Why? Customers often ask whether the Kitov has artificial intelligence and whether neural networks are available But what do these overused technical terms actually mean in detail? Even companies that deal intensively with this issue scrutinize algorithms and are familiar with AI raise the question of how AI should be interpreted For other tools there are still no accurate technical terms - Icall all of this “fuzzy” AI Asmall scratch on the surface of a component for example is impossible to assess automatically because it might affect functionality in one place yet simply be unsightly in another AI cannot make that judgement The Kitov however is equipped with tools developed in AI laboratories which is why it is possible to determine exactly which problems can occur where Moreover the Kitov can see everything that a human can see at least as well and even a little better When the machine inspects a product it compares it with a good sample or with stored image files just as a human would In addition again like a human the Kitov independently illuminates a spot until it achieves the best possible contrast In contrast to an AOI automated optical inspection with a static camera the Kitov can also detect hidden parts To do this it moves close to the position to be inspected and examines it from different perspectives It compares the image obtained in this way with CAD data libraries and stored instructions The device can thus assign categories and inform an operator about any deviation Moreover the Kitov learns from humans how to classify the material to be inspected into process indicators such as good bad or ‘somewhere in between’ The device thus has functions that are recognized as AI Other devices however are equipped with functions that can be classified as clever algorithms but do not meet the definition of AI In which applications is the Kitov One currently being used? In Germany the Kitov is now in use in the electronics automotive medical and computer technology sectors In other words everywhere where high quality is required For example one of our customers buys used PCs and servers in order to overhaul them Here the Kitov checks to what extent for example the keyboards and housings are still usable Another company uses the Kitov to automate the inspection area at an automobile manufacturer In the field of medicine the AIbased robot inspects tools used for the series production of pills The pills must be absolutely precisely shaped and fully intact At the same time they are produced at considerable speed which is why the tools are subject to high wear and tear But the Kitov can also inspect aircraft turbine blades or other intricate products But it can also be used in mobile applications and therefore inspect large products such as switching cabinets The range of product areas to be inspected is thus broad Nevertheless German companies often lack the courage to invest in forwardlooking technology If we want to secure the location we have to take up these new topics and implement such technologies This also requires a certain willingness to take risks and to evaluate We have already proven that the necessary investments pay off quickly The interview was conducted by Nicole Wörner ➡ continued from Page 1 “ AI is a long way from becoming mainstream”