DynamicBackgroundExtraction Sample Point Question

[jimwaters]

Is it possible to select too many sample points?  Is there an optimum number of sample points?  What about in areas with heavy nebula such as Orion?

Thx

[pfile]

it really depends on the image. if your gradients are nice and smooth and linear, then fewer points that capture that gradient well are preferable to a whole bunch.

if you have really complex gradients in your image, then more samples are necessary to capture all the variations in the gradient.

either way, try to keep the sample well away from stars, especially bright ones. while generally speaking the bright core of the star will be rejected by the algorithm, the halo around the star may be picked up and interpreted as background, which will then cause a false brightening of the extracted background in that vicinity.

heavy nebulosity is hard. for orion i usually just pick the 4-5 darkest areas of dust and go with that.

rob

[jimwaters]

Thanks Rob

[Warhen]

Hi Jim, I agree with Rob. I'd just add, that because gradients are generally large and smooth structures, they're relatively easy for DBE/ABE to detect. I tend to go 'less is more' as not to confuse their task, placing them in ideal positions, as Rob describes- avoiding halos of stars and galaxies, etc. Remember too, the power of the extended features such as Symmetry. Remember to use available clues to reveal neutral areas, such as studying a single channel (red or an Ha image when working on emission nebula).