Hi Emanuele,
Not necessarily. The ImageIntegration tool will automatically normalize and scale your images, as necessary. However, the LinearFit tool implements a more accurate fitting algorithm, which we haven't implemented in ImageIntegration for performance reasons.
If you can/want afford the extra processing effort (LinearFit is relatively slow), fitting all your images with LinearFit prior to integration (e.g. as a batch process with ImageContainer) might lead to a slightly better pixel rejection —in theory this is true; what I doubt is if doing so would actually lead to better results, in practice.
Let me know how it goes if you make a comparison (for example, with the same ImageIntegration parameters and with/without pre-normalization of the images with LinearFit).
