You would save the stretched values from the display. This outputs a byte-scaled image between 0-255 so if you need the original values for your calculation, you may not want to do this.

Thanks for the reply. No, I want to keep my original values. It would be handy to save the stretch in situ or at least export the stretch to a new file that can then be used for further analysis. I guess this would easier in IDL yes? Wes

Well, if you are forcing one image into the same data range as another, there will be changes in the output values of the 'matched' image. I might use something like Basic Tools ->Stretch Data to create two datasets with the same range. This tool will allow you to set the min/max values and data type you want for output. You could calculate the statistics on the first image used 'to match' and then set the same min/max and use a linear or other type of stretch to get the second 'matching' image to have the same range as the first image. You may still have a difference in illumination along the seam but it seems to me that your goal is less about making the images look similar than actually having their data ranges be similar.

Fantastic MariM, thank you so much. I am more interested in appropriate quantitative change detection methodologies than pretty pictures. I can make those in ArcGIS / QGIS :-) Just as a side note, I can probably use this very same tool to convert a MODIS MOD13Q1 NDVI data to the suitable -1 to 1 value range (instead of dividing by 10000)? Many thanks for your help.

Hi Support team I have the same problem. Even by stretching saved file of image which has been matched by the first base image, the float values of stretched file is not close to the base. It is very strange why ENVI does not allow saving files with the real float values after histogram matching and saves the output in 8 bit integer values. Is there any solution? Please help. Cheers Abbas