Yes, CFA is the abbreviation for Color Filter Array. Most DSLR and CMOS One Shot Color cameras have a Bayer matix of color filters in front of the sensor. In case of the widely-used RGGB matrix, this means alternating R and G filters or G and B filters in each row of pixels. A 4 x 4 pixel section of the sensor looks like this:
R G R G
G B G B
R G R G
G B G B
Each pixel of a raw image contains 1) the color information (i.e. either red, green or blue) and 2) an intensity. Nevertheless the raw CFA image (for a Canon DSLR camera: in the CR2 file) is monochrome (black and white). The color information is in the spatially separated pixels and has to be "developed" in order to generate a color image. This process, called debayering, is an interpolation. Assume that the sensor has w pixels in width and h pixels in height, making a number of total pixels of n = h * w. Then the sensor contains 0.25 * n pixels with red information, 0.5 * n with green information and 0.25 * n with blue information. The debayer alorithms interpolate the color information for all of the n pixels, and only after this "debayering", the CFA image is transformed into a RGB (color) image. If this description ist still unclear to you, I recommend to search the internet for the terms used here.
The preprocessing (calibration and cosmetic correction) can be performed on CFA images. This approach has the advantage of processing the raw (not interpolated) data and additionally of saving disk space for the intermediate files (master calibration files and calibrated light frames). After calibration (and optionally cosmetic correction) the debayering process is necessary, and not before this step a color image is generated.
Bernd
