Improving Automatic Surface Inspection Performance by Multiple Synchronized Views and Enhanced Classification Algorithms

Abstract

ABSTRACT The demand for high-grade rolled steel is steadily increasing and driven mainly by the automotive industry. Another key driver besides producing highest quality grades is to reduce the overall CO2 footprint by increasing the efficiency and yield of each process. To increase the throughput of each manufacturing step and reduce resource need, reliable data continuously gains importance, in which automatic surface inspection systems play an essential role. As automatic surface inspection is state-of-the-art, this paper details new approaches to increase the information quality provided by the inspection system. Results for multiple pixel-synchronized views and enhanced classification (neural networks) are presented.

Authors 1

1. Dominik Recker Aachen ISRA VISION Parsytec

   ISRA Vision Parsytec (Germany)

   Affiliation as printed

   ISRA VISION Parsytec, Pascalstraße 16, 52076 Aachen, Germany