| Zugriffsnummer | 14451 |
| Dokumenttyp | Konferenzartikel |
| Sprache | Englisch |
| Titel | Correlations in photometric measurements |
| Autor(in); Institution |
Sauter, Georg; 4.101, Photometrische Entwicklungen, PTB-Braunschweig
|
| Quelle/Jahr | Proceedings of the CIE Expert Symposium 2001 on Uncertainty Evaluation: Methods for analysis of uncertainties in optical radiation measurement:(2001), 18 - 22 |
| Schriftenreihe | CIE X: 020 |
| ISBN | 3-901906-05-3 |
| Verlag | Vienna: Commission International de l'Eclairage |
| Konferenzangaben | CIE Symposium 2001 on: Uncertainty Evaluation - Methods for Analysis of Uncertainty in Optical Radiation Measurement, Vienna, 23-24, January, 2001, Austria |
| Freie Schlagworte | uncertainty ; correlation ; monitor-technique ; interpolation ; integration |
| Zusammenfassung | Photometric measurements, which are influenced by correlation are discussed and the determination of a correlation coefficient is explained using statistical, logical and numerical methods. It is shown, how correlation of input quantities can increase or reduce the combined uncertainty associated to the output quantity and how to avoid the handling of correlation either by an additional quantity or by replacing input quantities e.g. by their ratios in "monitor-technique". Correlation between calibrated transfer standards created by a common reference standard is found in laboratories and also for key comparisons in the mutual recognition arrangement, MRA. Averaging of the measured values of such a batch of standards under-estimates the transfer uncertainty, except the results are used in normalised presentation. Measurements for the realisation of photometric units or for a calibration of the spectral power distribution of lamps include a numerical interpolation between function values, which creates correlation. Two examples are discussed, spectral responsivity and spectral distribution. The combined uncertainty for the integration of input values with a common positive correlation between all their pairs is calculated and the variation with the correlation is explained. |
