1.  ENVI will expect a single file with the VIS-SWIR bands present, so 6 bands for TM and ETM (1-5 and 7) and 4 bands for MSS. 2. For ETM data, the data should first be converted to reflectance.  For TM or MSS you can input the data in DN.

Thank you! Delving a bit deeper: Ideally I'd like to input reflectance data to the tasseled cap transformation - the Landsat-5 TM scene is just one in a 20-image timseries that includes TM and ETM data, which I've already corrected to reflectance (but of course the DN data are also available). So...is it okay to use Landsat-5 TM reflectance data, or must I use DN data? Many thanks!

The TC transform in ENVI for TM data is based on the values from the article mentioned in the help.  These are empirically derived values developed specifically to be used with DN.  Crist and Cicone came out a year later with a set of transformation values that could be used for TM reflectance data (A TM Tasseled Cap equivalent transformation for reflectance factor data.  CRIST, E P.  1985).  So I think it would not be appropriate to use ENVI's TC transform for TM with data converted to reflectance.  If you wish to use your reflectance data, then you should manually apply the equations to your reflectance data using the appropriate TC values. 

Thank you! This clarifies - I will manually apply the equations from Crist 1985 to the reflectance data.

Hello, I have a question. Is it possible to conduct a Tasseled Cap Transformation to an IKONOS image? In a paper by Horne, he develops the IKONOS Tasseled Cap coefficients. Is it possible to use these coefficients in ENVI and if so how. Thank you Colibri

ENVI does not have built-in coefficients for the TC Transform for IKONOS.  You would have to apply the equations manually using Band Math.

How would I do that? The formulas for the Tasseled Cap Coefficients for IKONOS are: TC1= 0.326 B + 0.509 G + 0.560 R + 0.567 NIR TC2= 0.311 B - 0.356 G - 0.325 R + 0.819 NIR TC3= 0.612 B - 0.312 G + 0.722 R - 0.081 NIR TC4= 0.650 B + 0.719 G - 0.243 R - 0.031 NIR And the vegetation index formula is: VI = a(TC2) - b(TC1) - c(TC3) , where a,b,c are weighing factors. Would I have to do this programming in IDL? Thank you Colibri

Hi Colibri, In order to set up the band math equation with the correct syntax you should refer to the ENVI Help system. If you use the Contents tab look here: Working with ENVI > ENVI Users Guide > Basic Tools > Band Math The explanations are fairly extensive and also provide numerous examples to guide you. To help you get started consider the following hint: B, G, R, NIR are your band variables. What I would suggest is to first try some basic band math equations on your image so you understand how this function works, and to make sure that the results are what you expect. Once you get the hang of it you can apply the TC equations you mentioned. Keep in mind that the ENVI Help system has a lot of information and is a good resource that shouldn't be overlooked. Ciao

Hi all, Crist 1985 specifically says that his tc transformation matrix was designed for landsat-4 tm and that it shouldn't be routinely applied to data from other sensors. However, it seems like many people do so anyways. Are there any studies proving the validity of this approach? Can the Crist 1985 matrix be applied to both ETM+ and landsat-5 TM?