TY - GEN

T1 - Surface inspection system for carriage parts

AU - Denkena, Berend

AU - Acker, Wolfram

PY - 2006/4/21

Y1 - 2006/4/21

N2 - Quality standards are very high in carriage manufacturing, due to the fact, that the visual quality impression is highly relevant for the purchase decision for the customer. In carriage parts even very small dents can be visible on the varnished and polished surface by observing reflections. The industrial demands are to detect these form errors on the unvarnished part. In order to meet the requirements, a stripe projection system for automatic recognition of waviness and form errors is introduced. It bases on a modified stripe projection method using a high resolution line scan camera. Particular emphasis is put on achieving a short measuring time and a high resolution in depth, aiming at a reliable automatic recognition of dents and waviness of 10 μm on large curved surfaces of approximately l m width. The resulting point cloud needs to be filtered in order to detect dents. Therefore a spatial filtering technique is used. This works well on smoothly curved surfaces, if frequency parameters are well defined. On more complex parts like mudguards the method is restricted by the fact that frequencies near the define dent frequencies occur within the surface as well. To allow analysis of complex parts, the system is currently extended by including 3D CAD models into the process of inspection. For smoothly curved surfaces, the measuring speed of the prototype is mainly limited by the amount of light produced by the stripe projector. For complex surfaces the measuring speed is limited by the time consuming matching process. Currently, the development focuses on the improvement of the measuring speed.

AB - Quality standards are very high in carriage manufacturing, due to the fact, that the visual quality impression is highly relevant for the purchase decision for the customer. In carriage parts even very small dents can be visible on the varnished and polished surface by observing reflections. The industrial demands are to detect these form errors on the unvarnished part. In order to meet the requirements, a stripe projection system for automatic recognition of waviness and form errors is introduced. It bases on a modified stripe projection method using a high resolution line scan camera. Particular emphasis is put on achieving a short measuring time and a high resolution in depth, aiming at a reliable automatic recognition of dents and waviness of 10 μm on large curved surfaces of approximately l m width. The resulting point cloud needs to be filtered in order to detect dents. Therefore a spatial filtering technique is used. This works well on smoothly curved surfaces, if frequency parameters are well defined. On more complex parts like mudguards the method is restricted by the fact that frequencies near the define dent frequencies occur within the surface as well. To allow analysis of complex parts, the system is currently extended by including 3D CAD models into the process of inspection. For smoothly curved surfaces, the measuring speed of the prototype is mainly limited by the amount of light produced by the stripe projector. For complex surfaces the measuring speed is limited by the time consuming matching process. Currently, the development focuses on the improvement of the measuring speed.

KW - Carriage parts

KW - Optical measurement

KW - Quality assurance

KW - Stripe projection

UR - http://www.scopus.com/inward/record.url?scp=33747712099&partnerID=8YFLogxK

U2 - 10.1117/12.662572

DO - 10.1117/12.662572

M3 - Conference contribution

AN - SCOPUS:33747712099

SN - 0819462543

SN - 9780819462541

T3 - Proceedings of SPIE - The International Society for Optical Engineering

BT - Photonics in the Automobile II

PB - SPIE

CY - Berlin, Bellingham, Wash.

T2 - SPIE Photonics Europe

Y2 - 3 April 2006 through 7 April 2006

ER -