Dark#5 - Average
binations of dark frames having integration times that are longer than the longest integration time used for any raw image. In calibration, a master bias frame is subtracted from the master dark, and the resulting thermal frame is scaled.
Many observers prefer the standard protocol because they do not wish to expend time and effort taking bias frames, and the standard calibration procedure is simpler. On the other hand, it forces the observer to use the same integration time for all images, or to shoot different sets of dark frames having different integration times.
However, the time required to make, for example, 16 bias frames and 10 dark frames with 300-second integration times for the advanced calibration protocol is repaid by precise subtraction of dark current with somewhat lower thermal noise. Moreover, for CCD cameras that produce electronically "clean" images, the master bias frame can be replaced with a single bias value, making the advanced calibration protocol just as easy to use as the standard procedure, but having the advantage of greater flexibility in choosing image integration times.
Because the accumulation of thermal electrons during integration is proportional to integration time, the number of thermal electrons in a thermal frame, i.e., a bias-subtracted dark frame, can be matched to an image with a different integration time simply by multiplying by the ratio of exposure times:
Knowing the integration times, a master dark frame made with 300-second integrations can be scaled to match to images with 60-second integrations. This is done by subtracting the bias value, multiplying the pixel values in the master dark by a factor of and then subtracting the scaled master dark frame.
Because of a small change in the temperature of the CCD, you may not know how many thermal electrons have accumulated, but it is possible to determine the relationship using the small population of pixels with higher dark current than most photosites—the so-called "hot" pixels. The dark-matching algorithm searches the dark frame for a representative sample of "hot" pixels and measures their excess compared to adjacent photosites, and then determines the excess for corresponding pixels in a raw image. For each hot pixel, the algorithm determines the ratio of the excesses, discards spurious values, and averages the representative ratios to obtain a scaling factor. The dark frame is then bias subtracted, scaled using that ratio, and subtracted from the image.
Was this article helpful?