Correction of measuring data of a spectrograph by using neural networks

Klaus Weber $^{\ast\ast}$

$^{\ast}$ Technische Universität Cottbus, Lehrstuhl Numerische und Angewandte Mathematik

Zusammenfassung:

The possibility of using neural networks in industrial practice is shown from the correction of measuring data of a spectrograph. First all steps are represented, from the description of the principle of measurment, the measuring instrument and the origin of the metering error, to the different neural resolving approaches. As the main point it is shown, how knowledge on the technical system can be used for pre-structuring of a neural network and how the performance of the latter can be increased by doing so clearly.

Quellennachweise:

KLAUS WEBER (1994).: Correction of measuring data of a spectrograph by using neural networks
In: 3. GI/ITG/GME-Fachtagung Rechnergestützter Entwurf und Architektur mikroelektronischer Systeme