No, there is no way to orthorectify your image in ENVI without appropriate metadata or RPCs with or without Ground Control.  If there are registration issues with your image and other data layers, you could perform an Image to Image Registration with another image or vector layer that is known to be accurate to try and improve the positional accuracy of your IKONOS image.

Thank you. I am getting the RPC file from the vendor and I have a 30m DEM. Is it ok to use a 30m DEM when my image is a 1m pan-sharpened IKONOS image? Also is there someway to interpolate the DEM over the pan sharpened image to better see the rivers (lowest elevations) when I am digitizing? Colibri

It is fine to use a 30m dem with your data.  If you are in a very mountainous area, a higher resolution DEM would be more accurate for capturing the variability in elevations, but it is not necessary.  Are you asking if resampling the DEM to the same pixel size as the IKONOS image would provide more detail?  I so, I don't believe it will.

Hello mminari, Thank you for helping so much on this forum! Ok. So My study area is a super mountainous area but the 30m DEM is the best I can do right now. So, it will have to do. I was asking if there is any visual way to use the contours of the DEM image to help digitize rivers in the pan-sharpened image. Some times in the pan-sharpened image the river is "lost" because there is allot of vegetation covering the water and rocks. However, it is very important to me to be able to "see" the rivers in the pan-sharpened image. I basically would like to create a vector layer of the rivers based on the DEM image (darker areas are lower and most of them have rivers...can I create a vector of the low elevation seen in the DEM? If so how? I do have a question about re-sampling but it is with my pan-sharpened image. I would like to do a number of classifications with different pixel resolutions (1m, 10m, 20m, 50m). My analysis requires that I see the vegetation types at different scales. How can I re-sample a 1m pan-sharpened image into the 10m, 20m, and 50m resolution? Appreciate your help Colibri

Hi Colibri, Open the DEM then from the Image menu chose Overlay > Contours When the next dialog opens chose Options > Add New Levels and then chose a Start level and an increment you wish to use and the number of levels you  want. If the image is large take a subset and experiment with setting till you like what you see. You should be able to apply that to the bigger image. This should allow you to do a reasonable job of setting up to contours to find your water courses. To save the contours as an ENVI vector file (*.evf) use File > Output Contours to EVF... If you have a good pan image you may be able to use differences in vegetation to determine where the water is flowing in areas where you cannot really see the water. Take a look at where you see water and check the texture of the tree canopy. You may find similar texture in areas where water and rocks aren't so visible. To resample try Basic Tools > Resize Data (Spatial/Spectral) Ciao

Super! I will try that. Thanks Colibri

Hello, I am going to conduct an unsupervised classification on a pan-sharpened IKONOS image, however I am preprocessing the image first and I have a few questions about the order of the pre-processing procedure. Is this the order of the preprocessing procedure?: 1. orthorectify each band (R,G,B,NIR,Pan) 2. IKONOS Calibration (DN to radiance) 3. Layer stacking (B,G,R,NIR) 4. Convert Data Tool (from BSQ to BIL) 5. Edit ENVI Header (with each bands coefficient) 6. FLAASH (atmospheric correction) 7. Apply ROI Tool 8. Gram-Schmidt Spectral Sharpening Tool 9. Masking I am not interested in and than I can conduct an IsoData Unsupervised Classification Thank you for your help Colibri

I would do it in this order: 1.  IKONOS Calibration to radiance (multispectral and pan, separately) 2. Convert multispectral image from BSQ to BIL 3. FLAASH correction to reflectance (multispectral and pan, separately) 4. Gram-Schmidt Spectral Sharpening (pan sharpening) 5. Unsupervised classification I don't think you need to do any layer stacking.  The B, G, R, and NIR bands should have the same orientation and resolution already, so even if you did separate them for some reason, you could put them back together using File > Save File As > ENVI Standard.  Layer Stacking is meant for cases where you need to do some processing to get the images to line up properly, such as when they have different map projections or pixel sizes. Also, if you are doing unsupervised classification, I can't see why you would need to use the ROI Tool.  You only use ROIs to identify training data for a classification if it is supervised. I also don't see a reason to edit the ENVI Header of your files.  But maybe I am missing something?  Was this to add the wavelengths for each band?  If you don't pull apart your multispectral image into separate bands, then you won't need to specify wavelengths when you put it back together.  ENVI can automatically assign the correct wavelengths to IKONOS multispectral images when it first reads the file. I hope this helps. Peg