Procedures

Two data components are transmitted to the ground base.  The largest data component is 
the compressed imaging spectrophotometer (ISP) and GPS data at XXX Mhz.  These will be decompressed
losslessly and stored with the time data for later tomographic analysis. The other data com-
ponent will be the attitude video camera (AVC) image received at XX Mhz. 

The AVC data with time data will first be decoded frame by frame to determine the direc-
tion in which the ISP was looking at each moment. This technique has already been demon-
strated successfully in Frank Harris' Gemini Project with the NRC.  The process involves 
computing centroids of the brightest regions in the frame, joining the centroids, computing 
the combination of angles, and matching these angles with a database of known star geo-
metry.  From these, the look 

Next, the ISP images will be processed to obtain the species responsible for stellar occulta-
tion and the corresponding column densities. The first step is to remove the blurring caused
by ISP movement during optical integration. This blurring will be primarily perpendicular to
the spectral data.  Therefore, adjacent spectra can be shifted and summed to compensate
for ISP movement.  This is accomplished by referring to the angular deflections obtained 
over time from the AVC results. This summation process will simultaneously intensify the 
spectra obtained and cause random noise present on each spectrum to cancel itself.

The cleaned up and intensified stellar spectrum will be compared with the unocculted spec-
trum for the star obtained before the line of sight crossed into the atmosphere below 40 km
altitude.  The ratio of these spectra will yield a wavelength-dependent transfer function which
may be decomposed into its constituent parts based on the known transfer functions of the
dominant atmospheric species above 10 km altitude: O3, NO3, and NO2.  The weights of
these component transfer functions will correspond to the constituent column densities of 
these species.

Each ozone column density furnishes a single linear equation linking the ozone concentrations in each
of the regions through which the starlight passed. The accumulation of all such equations represents
a linear system awaiting solution by a known tomographic algorithm for the ozone distribution in the 
vicinity of the instrument's trajectory. Because the deviation of the columns from the horizontal is very 
slight, the information about the vertical ozone distribution will be much richer than for geographic 
distribution.  Correspondingly, the vertical resolution of (expected to be +/- 1 km) will be much better 
than the geographical resolution (+/- 100 km). - John Steele