SMEI consists of three cameras, each with a clear field of view of roughly 60°×3°, aligned in such a way on the spacecraft that combined they image a strip of sky of about 160°×3°. As the spacecraft moves through its polar orbit, the SMEI cameras continuously take 4-second exposures and sweep nearly the entire sky over each 102-minute orbit. CCD frames from each orbit (typically 1500 from each camera) are combined to provide a photometric white-light map of the sky as visible from Earth orbit. In the analysis done at UCSD, point sources are removed from these composite skymaps, rather than from the individual CCD frames. A typical sky location transits the field of view at an approximately fixed position in its long dimension, crossing the 3° narrow dimension in about one minute. Thus, about a dozen or more sequential CCD frames contribute to the brightness at that location in the composite skymap.
All frames collected over a single orbit (about 1500 for each camera) are combined to form one set of full-sky maps, in coordinate frames as in the accompanying video. The assembly is achieved using a cartesian grid with an angular resolution of about 0.1°. The grid covers the celestial sphere in rectangular bins that within acceptable limits, about a factor of two, have the same area, and hence the bins do not suffer from the strong edge deformations inherent in common planar map projections of the celestial sphere. All pixels within the acceptable field of view of all frames are "indexed" by placing them in the appropriate bins in the grid. Each CCD pixel of about 0.05° on the sky in engineering mode, and 0.2° in science mode, contributes to several bins on the grid. Each sidereal location in the sky (i.e., each bin on the grid) receives contributions from a sequence of subsequent frames recorded as the camera field of view sweeps across the sidereal location. This combination of contributions from the time domain provides a fortuitous but elegant means of detecting temporally isolated spikes in the CCD frames, such as particle hits (isolated high values in one or more pixels in a single frame) and space debris (traveling across the field of view in times short compared to the exposure time). Thus, in addition to preparing the full-sky sidereal maps, the indexing phase also is our main defense against these short-lived contaminations in the data.
After the 0.1° resolution grid is produced, it is reformatted to a lower-resolution set of sidereal maps of sky surface brightness. From these sidereal maps we remove bright stars, background stars, and a zodiacal cloud model (their brightnesses are retained as additional data products). The final maps can be represented in any convenient sky coordinate system. Common formats are the above RA vs declination, Sun-centered Hammer-Aitoff, or "fisheye" maps. Time series at selected locations on these maps are extracted and processed further to remove aurorae, variable stars and other unwanted signals. These time series (with a long-term base removed) are used in 3D tomographic reconstructions.