The noise power spectrum, NPS, is a key imaging property of a detector and one of the principle quantities needed to compute the detective quantum efficiency. NPS is measured by computing the Fourier transform of flat field images. Different measurement methods are investigated and evaluated with images obtained from an amorphous silicon flat panel x-ray imaging detector. First, the influence of fixed pattern structures is minimized by appropriate background corrections. For a given data set the effect of using different types of windowing functions is studied. Also different window sizes and amounts of overlap between windows are evaluated and compared to theoretical predictions. Results indicate that measurement error is minimized when applying overlapping Hanning windows on the raw data. Finally it is shown that radial averaging is a useful method of reducing the two-dimensional noise power spectrum to one dimension.

Date Published: 28 June 2001
PDF: 11 pages
Proc. SPIE 4320, Medical Imaging 2001: Physics of Medical Imaging, (28 June 2001); doi: 10.1117/12.430934

Published in SPIE Proceedings Vol. 4320:
Medical Imaging 2001: Physics of Medical Imaging
Larry E. Antonuk; Martin Joel Yaffe, Editor(s)