Share Email Print
cover

Proceedings Paper

Optical classification for quality and defect analysis of train brakes
Author(s): Stefan Glock; Stefan Hausmann; Sebastian Gerke; Alexander Warok; Peter Spiess; Stefan Witte; Volker Lohweg
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

In this paper we present an optical measurement system approach for quality analysis of brakes which are used in high-speed trains. The brakes consist of the so called brake discs and pads. In a deceleration process the discs will be heated up to 500°C. The quality measure is based on the fact that the heated brake discs should not generate hot spots inside the brake material. Instead, the brake disc should be heated homogeneously by the deceleration. Therefore, it makes sense to analyze the number of hot spots and their relative gradients to create a quality measure for train brakes. In this contribution we present a new approach for a quality measurement system which is based on an image analysis and classification of infra-red based heat images. Brake images which are represented in pseudo-color are first transformed in a linear grayscale space by a hue-saturation-intensity (HSI) space. This transform is necessary for the following gradient analysis which is based on gray scale gradient filters. Furthermore, different features based on Haralick's measures are generated from the gray scale and gradient images. A following Fuzzy-Pattern-Classifier is used for the classification of good and bad brakes. It has to be pointed out that the classifier returns a score value for each brake which is between 0 and 100% good quality. This fact guarantees that not only good and bad bakes can be distinguished, but also their quality can be labeled. The results show that all critical thermal patterns of train brakes can be sensed and verified.

Paper Details

Date Published: 17 June 2009
PDF: 9 pages
Proc. SPIE 7389, Optical Measurement Systems for Industrial Inspection VI, 73891Y (17 June 2009); doi: 10.1117/12.827457
Show Author Affiliations
Stefan Glock, Ostwestfalen-Lippe Univ. of Applied Sciences (Germany)
Stefan Hausmann, Ostwestfalen-Lippe Univ. of Applied Sciences (Germany)
Sebastian Gerke, Ostwestfalen-Lippe Univ. of Applied Sciences (Germany)
Alexander Warok, Ostwestfalen-Lippe Univ. of Applied Sciences (Germany)
Peter Spiess, Deutsche Bahn AG (Germany)
Stefan Witte, Ostwestfalen-Lippe Univ. of Applied Sciences (Germany)
Volker Lohweg, Ostwestfalen-Lippe Univ. of Applied Sciences (Germany)


Published in SPIE Proceedings Vol. 7389:
Optical Measurement Systems for Industrial Inspection VI
Peter H. Lehmann, Editor(s)

© SPIE. Terms of Use
Back to Top