Exposure Correction Factor The length of exposure recorded in the image header has some error. This is corrected by multiplying nominal exposure times by an exposure factor which is generated as described in this document. The main exposure factor procedure, compute_monexp_factors, computes exposure time correction factors of full-field images by comparing the median intensities in various regions. A region is a group of 32x32 superpixels. The ratio of a given image to a reference image is computed. The median of the median ratios of all the superpixels in a region is computed. The time history of the region medians is used to detrend (each region separately) the normal coronal variation from the exposure time fluctuations. Then the average of all the detrended region ratios is computed to form the overall correction factor for each image. The detrending fits the 11 images before and after the time of the desired image to a 2nd degree polynomial. No correction is made for irregular timings. All sub-images, plus the full-field images that could not be corrected by the main routine, are passed to another routine, corrsubimage, for interpolation/extrapolation. The following color code is used to distinguish such images in these plots. To interpolate/extrapolate, only full-field images for which the main routine has already calculated valid exposure factors are used. Images earlier and later in time to the input image are tested to see if they have superpixel blocks that overlap the input image. Two images closest in time to the input image are selected and ratios of overlap regions of each image to the input image are calculated and averaged. Then a simple linear interpolation between the two times is performed to determine the exposure factor for the input image. The standard deviations for the exposure factors of the interpolated images, plotted in green, and the extrapolated images, plotted in yellow, are set to ones by the corrsubimage routine. There are times that corrsubimage routine can not locate good images to interpolate/extrapolate. In these cases, the exposure factors and standard deviations determined by the main routine (1.0 and 0.0 for sub- images) are returned. These are plotted in red. Finally, some images have invalid roll angles. Corrsubimage sets the exposure factors and standard deviations for such images to ones and zeros, respectively. These are plotted in purple. And, when the exposure factors still seem to be incorrect (such as the proton storm event of November 2001 in C2), these exposure factors and standard deviations are also set to ones and zeros to indicate this fact. The exposure factors for these images are plotted in blue. The ones and zeros as standard deviations for the green, yellow, purple, and blue exposure factors are only meant to distinguish the corrsubimage calculated factors from those determined by the main routine (plotted in white). Unlike the derived standard deviations in the main routine, these standard deviations of ones and zeros are meaningless and must be ignored. Each of the following png files contains monthly exposure time correction factors and corresponding standard deviations plots for C2 (normal, orange, clear) or C3 (normal, clear, clear) images. Monthly C2 and C3 Exposure Factor Plots are grouped by year (YYYY). Files are named by telescope (c2 or c3) and two-digit year and month (YYMM). ie. YEAR Month Tel File 1996 Jan C2 c2_9601_lz_efac_std.png 1996 Jan C3 c3_9601_lz_efac_std.png Exposure factor plots are included for all months from January 1996 to December 2005 except July 1998 to September 1998 and January 1999. ------------------------------------------------------------------------ Downloaded: http://lasco-www.nrl.navy.mil/expfac/expfac.html Modified: Matthew Knight 16 Feb 2004; 15 Mar 2006; 27 Jan 2008 Puneet Khetarpal 17 Aug 2004; 07 Feb 2005; 25 Mar 2005; 12 Apr 2005 Trishla Maru 01 September 2005