Science.Online
Publisher and Institutes
Akademie Verlag
Deutsches Institut für Urbanistik
Oldenbourg Wissenschaftsverlag
Walter de Gruyter
Schattauer
You are here: Home :: Area NEM :: Engineering :: Machine construction
 
P. Lehmann, S. Patzelt, Gert Goch

Surface Roughness Characterization based on Polychromatic Far-field Speckles of Continuous or Discrete Spectral Distribution

It is well known that polychromatic light scattering is suitable for a roughness characterization of engineered surfaces even if the height differences exceed the wavelength of light, so that conventional light scattering techniques fail. However, polychromatic light can be realized by both, a combination of several laser sources emitting at different discrete wavelengths or a single source emitting continuously distributed light of a narrow spectral bandwidth.
In this contribution, firstly a modified definition of the term coherence length is introduced in order to compare the physical consequences of these two realizations. Within the context of roughness characterization the phenomenon of interest is the speckle elongation which can be obtained from a single black and white CCD-shot of a far-field speckle pattern by autocorrelation analysis. This is the reason why this method is suitable for in-process measurement on rapidly moving objects.
It is shown that – with the same coherence length – continuously distributed polychromatic light, as it is emitted by a super luminescent diode, for example, offers fundamental advantages. Experimental results have been obtained using discrete (trichromatic) as well as continuously distributed polychromatic light. The spectral bandwidth and the surface roughness have been varied in both cases. The results confirm the theory and demonstrate the advantages of light sources with a continuous spectrum for this application.

tm – Technisches Messen, Oldenbourg Wissenschaftsverlag

Print ISSN: 0171-8096
Volume: 69, 05/2002
Pages: 263

Journal homepage (external site)

Show all available items of this journal