PDS_VERSION_ID = PDS3 RECORD_TYPE = STREAM OBJECT = TEXT PUBLICATION_DATE = 2007-05-25 NOTE = "Description of CTX calibration process" END_OBJECT = TEXT END The following explanation of the CTX calibration process assumes 0-relative indexing; e.g. the first pixel of any given line has index 0. File lines and dataset lines are also numbered starting with 0. There are two input ancillary data files required for this process: the decompanding table (ctxdec.txt) and the flattening table (ctxflat.txt). Both of these files are simple ASCII files. 0) Read the decompanding table file, ctxdec.txt. ctxdec.txt contains 256 lines, one line for each possible byte value [0, 255]. There is one value per line, namely the decompanded value corresponding to the line number. 1) Read the flattening table file, ctxflat.txt. ctxflat.txt contains 5064 lines, 8 more than 5056, the number of pixels per sum=1 full length CTX line. The last 8 lines of ctxflat.txt are never used. There are two values per line: pixel index followed by a flat divisor value. 2) Align the flattening table with the given CTX line. This is only necessary when the event start is non-zero. In this case, the flat table value with index (event start - 16) corresponds to the first pixel on the line. 3) If the event summing is greater than one, average the aligned flattening table to matching the event summing. For example, if the event summing is two, flat[i] = (flat[2*i] + flat[2*i +1])/2. 4) Compute the average even dark pixel value and the average odd dark pixel value. This can be done using the 'current' line, using some subset of lines bracketing the current line, or using all the lines in the dataset. The first 16 pixels of each sum=1 line are the dark pixels; they must be decompanded before averaging. The average even dark pixel is based only on pixels with even indexes; the average odd pixel is based only on pixels with odd indexes. Note: Dark pixel with index 14 appears to always run high. It may be prudent to avoid including this pixel in the averaging process. Also, for sum=2 data, only the first 8 pixels of each line are the dark pixels. 5) For any given line, decompand each pixel with even index, and subtract off the average even dark value. Decompand each pixel with odd index, and subtract off the average odd dark value. 6) Apply flattening to the decompanded CTX line, pixel by pixel. For each pixel index i, if flat[i] is 0, output 0, otherwise output pixel[i]/flat[i]. 7) Convert the flattened data to radiance I/F using the following formula: I = DN / Exposure Time / Summing / Response Coefficient F = Solar Irradiance / Pi / Solar Distance^2 where DN is the flattened value Response Coefficient = 8.55 +/-1.84 [units: (DN/msec) / (W/m^2/micron/sr)] Solar Irradiance Value = 1690 +/-5 [units: W/m^2/micron] (integrated over the CTX bandpass, at 1 AU) Exposure Time is in milliseconds Solar Distance is in Astronomical Units (AU)