Hi Siegfried,
there are two main areas in the script interface you must understand: "Intensity Maps" and "Composition Mask".
Intensity Maps[/size]
My algorithm works putting the shorter exposures over the saturated areas of the longer ones. For doing this, we must fit all the exposures; this means a background substraction and scaling of the images. The goal is to have the same intensity values in all the images for a certain object. This is done through the intensity maps.
An intensity map in a set of pixels between an intensity range. To calculate the scaling of target image, we select in a reference image two different intensity maps. One intensity map has higher illumination level than the other (in the script, the default intensity maps are between [0.8 - 0.6] and [0.4 - 0.2]. After selecting the pixel sets, you calculate the average illumination in the two maps (in this case, this will be nearly 0.7 and 0.3).
The key is that, if you compare the average illumination of these pixels with the average illumination of the same pixels of the target image, you know how much light intensity has the target image in comparison to the reference. In other words, you are looking how high are the highlights and how low are the lowlights in the target image, in comparisonto the reference one. This gives you the scaling factor of the target image.
The default intensity ranges will work in most of cases. But you must be sure you have pixels within this ranges in all the image set.
Composition Mask[/size]
As I said, my algorithm works putting the shorter exposures over the saturated areas of the longer ones. This is done through a mask, that selects a limit illumination to put the shorter exposures. The default value, 0.8, is a conservative one, to avoid the non-linearity area of the image sensor. This means that, in the pixels where the illumination is over 0.8 in the longer exposure, the shorter one will be superposed.
You have also a smoothing of the mask. This is necessary to prevent artifacts in the edges of the mask when compositing the images. This artifacts are a product of calculation error of the scaling of the images (due to intrinsic image noise). Usually a value between 1 - 3 will do the work.
The parameter mask growth will do a growth of the mask. This is useful to get rid of bloomings.
I hope this will help you... remember, you need very good image calibration, specially for the flats! If you get an image plenty of artifacts, surely it will be due to your image calibration, not to the algorithm. :wink:
Best regards,
Vicent.[/i]
