Dear F   Thank you ver much for your message. I confess that this is the first time I try to radiometrically or atmospherically correct a group of images. So any advice you can offer would be very much appreciated. I first tried to use Empirical line to calibrate the 2 Spot 5 images and 2 of the Spot 2 images, to the remaining Spot 2 image. This, however, didn't changed the Spot 2 images considerably, so they remain with different color tones to the base Spot 2 images. On the other hand the Spot 5 images (which have an extra SWIR band), changed completely in color, not correctly representing the area (I think). I also tried a dark object substraction, to see if I could bring all the images to substrate reflectance, but I don't see much difference in them after the analysis. What exactly should I be expecting?. I am particularly concern with one of my SPOT 2 images which appears to be much brighter than the other 2 Spot 2 images. All 3 of them were taken on the same day, but due to this difference I can not analyse them all at once in a mosaic, cause this will considerably affect my analysis. I am also wondering how can i calibrate the extra band of the Spot 5 images if I use a Spot 2 image as base image (like in the Empirical Line process), or is it better to go for something like the dark object subtraction that bring all images to a base line that is not dependent on an image per se. In addition,is there a way in which I can verify that the correction has been carried out properly? It seems to be that I am doing this with no estimation of the error, and no proper criteria to decide if the results are what I was expecting. Can you also suggest the best way to go to correct the images so I can put them all in a mosaic and performed by supervised classification? Right know I am tempted to classify them independently, but this will be much more work. In relation to your question about the format of my Spot imagery, these are 1B level Spot images. as far as I can see the format is 8bit unsigned interger, so I am guessing radiance levels and no Digital number, but I might be wrong. Minimum = 0 and Maximum =254.   Thank you in advance for your help   S

Dear F   Thank you very much for this information, this has certainly help me a lot.   A final question. I have decided to classify my images independently to avoid the problems with the radiometric correction. I intend to use, however, texture indices and a NDVI as ancialliary data for the classification, I wonder if this will still requiere the radiometric correction eventhough the classification will be done for each image separately, and the indices will be calculated and used only in reference to the original image?   Thanks again   S

Dear F   Thank you very much for your reply, this certainly helps me with the analysis.   S