| Author | Message | ||
     ErSk |
Hello Fellow NIR people I would like to learn how to transform my NIR spectra into polar coordinates ?? Could someone please help me with that ?? meaby a reference or email me a paper/description ([email protected]) Also I would like to hear about your experience with PQS. Thank you very much | ||
| David W. Hopkins |
The technique has been presented by Karoly Kaffka at several International Diffuse Reflectance Conferences (2000, 1998 and 1996 at least, I'd guess). The basic idea of transformation from Euclidean (x, y) space to polar coordinates (r, t) is the same as discussed in many geometry textbooks. However, in practice, Kaffka has described three methods for accomplishing the transformation. He has a Windows-based program for the tranformations and use for qualitative data analysis. The software is available directly from his company, Metrika, or at www.chemometrics.com. I'm sure you could write him at: Prof. Karoly J. Kaffka University of Horticulture & Food Industry Menesi ut 45 1118-Budapest, Hungary phone 011-26-1-166 46 35 e-mail [email protected] For references you could start by looking at the Proceedings of the IDRC for some years. | ||
| hlmark |
David's response will guide you to a good deal of the relevant information. There is also a company named Metrika now that is selling software that implements the approach, and includes a good deal of advanced analysis. The short answer to the original question about the conversion to polar coordinates, however is as follows: You start with a spectrum containing n data points, indexed 0 to (n-1), regardless of the wavelengths those correspond to. Let i be the index of any given measured wavelength in the spectrum, also having values 0 to (n-1) Then theta (i) = 2 * pi * i / n and R(i) equals the value of the absorbance of the spectrum at the wavelength with index i. Howard |