Neural Network-Based Vision and Signal System for Industrial Quality Control


Keywords DSPs, image processing, inspection systems, neural networks, vision systems

Start Date: 01 August 1992 / Duration: 24 months / Status: ongoing

[ contact / participants ]

a modular and flexible digital signal processor based vision and sensor signal system incorporating neural networks for real-time quality control applications

Objectives and Approach

A modular hardware platform will be developed consisting of digital signal processor boards for the host. This platform will interface to both matrix and line-scan cameras and to other sensors. A decision-making unit will also be implemented. The main emphasis will be on neural network algorithms, together with a tool for designing complex application systems.

The quality-control platform will be tested and evaluated in three demonstrator applications. One of these concerns the quality control of natural products (i.e. wood surfaces) and the two others concern the quality control of man-made products (i.e. potentiometer curves and car engines). The former will use vision-based sensing, while the two others will use signals from other sensory modes.


Christian Christiansen
Kjærgård Industri Automatic A/S
Grundtvigsvej 6-8
DK-8723 L»sning
tel: +45 7565 0000
fax: +45 75650199


Kjærgård Industri Automatic A/S [DK]
Hema Elektronik -Fertigungs und Vertriebs GmbH [D]
Mimetics SA [F]

HPCN synopses home page HPCN acronym index HPCN number index
All synopses home page all acronyms index all numbers index

NEUROQUACS - 7185, December 1993

please address enquiries to the ESPRIT Information Desk

html version of synopsis by Nick Cook