The Solar Oscillations Investigation project is a program at Stanford University to visualize and record the dynamics of sound wave propagation within the Sun. The dynamics of solar sound waves, and their effect on the surface motion of the Sun, can tell astronomers much about the interior structure of the Sun that cannot be observed with light alone.
The SOI project will use a Michelson Doppler Imager (MDI) instrument to detect solar surface motion caused by sound propagation. This instrument will be flown aboard the the Solar and Heliospheric Observatory (SOI) spacecraft, a joint NASA-ESA venture, which is scheduled to be launched from Florida in early December.
Before fitting for launch, the instrument was tested and calibrated with ground-based observations. During the test process, a problem with the low-frequency range of the CCD detectors was discovered and corrected. As part of this process, digital images of the Sun, taken with a CCD camera, were processed to eliminate the low-frequency anomaly. The following is a discussion of the technique used to process these early images.
Many thanks to Philip Scherrer for the data used in this analysis. All computations and displays below were performed with the Galactomatic-1000 (TM) scientific data analysis and visualization software product for the Macintosh. See Page 3 for a list of references and links.
The raw instrument image is a 256x256 pixel 32-bit FITS format file which has been obtained from a CCD camera in the SOI instrument. This image is just one out of hundreds (or more) of frames which will be combined to create a time-lapse movie of the sun. The low-frequency anomaly can be seen as a light patch in the lower left quadrant of the image. The goal of this analysis was to remove this anomaly without disturbing the rest of the data, and to develop a process which could then be applied to all other frames of the data.
The first step in removing the anomaly is to create a 2d Fast Fourier Transform of the raw image. The display above is the magnitude of the transform of this image. An FFT is a complex-valued result, so that the magnitude shown above is actually
where real(FFT) and imag(FFT) are the real and imaginary parts of the transformed image. In the display, low frequencies are in the center, with higher frequencies (positive and negative) away from the center.