Smart Industry

Moderator:

11:40

Industry 4.0 Human Interface Mate as a virtual AR solution

Arkite NV aids the Industry 4.0 uprise. with the Human Interface Mate, a hard and software tool that assists operators in making less mistakes easy in production.  Focus for Arkite is on autonomation (one of the key pillars of Lean management) instead of automation because people matter.

Ives De Saeger, Managing Director - Arkite
12:00

How Visual Positioning Systems augment our reality and robots

Intermodalics, a robotics software development company in Heverlee, Belgium has worked closely with Google for more than 2 years to develop a Visual Positioning System called Tango. It's a beaconless positioning system that works in any environment we can think of. Nowadays, it serves as the founding technology for ARCore, an Augmented Reality (AR) toolkit for Android, but it harbors great potential for Virtual Reality and Robotics applications that will influence our lives on a daily basis. This presentation gives an overview of the possible applications in industry, robotics and consumer markets: augmenting position information in GPS denied environments,measuring 3D shapes without 3D cameras, enabling true human-computer interactions within our living environment.

Peter Soetens, Director - Intermodalics
12:20

Business Model Innovation through data

Industry 4.0 delivers huge potential for industrial companies. Currently however, most Industry 4.0 data are not used. That's because this information is used mostly to detect and control anomalies - not for optimization, prediction and uptime, which provide the greatest value.

By collecting the right data at the right moment, the right decision will be made to transform and innovate your business model (servitization), helping you towards a connected Industry 4.0.

Jo Nelissen, CEO & Founder - SmartLog
12:40

Applied neural network classifiers in industry applications

Vision defect detection normally consist of aligning a product image with a common reference followed by finding pixel regions which deviate from the expected value. For products with highly varying shapes and texture such as flowers but also metal casted parts this approach is not feasible. The last decade neural network and deep learning approaches are becoming popular tools for product classification and defect detection. These methods work on training an application with large amounts of usually manually labeled images. After which the application can automatically classify new images into the trained categories. However; Trained classifiers are highly sensitive and will fail if illumination, location or the product changes in time.  Also manually labeling takes often too long and contains errors. Thus in practice many challenges has to be dealt with for a useable, efficient, maintainable industrial application is obtained

Dirk-Jan Kroon, Senior Vision Engineer - DEMCON

Ready for the SSIS 2017 Conferences?

Register as a visitor