What projection is your 5m HSI dataset in? It probably is something other than geographic lat/lon so you could try reprojecting that to your 5m image. What was the problem you were encountering with the crosshairs and the 2 images? With respect to the difference in albedo; don't forget that you are comparing a 500m pixel to a 5m pixel, the sampling area is quite a bit different, 250000m^2 versus 25m^2. The latter is much more accurate for a given area than the former.

Apologies if my post wasn't clear. The difference in albedo issue was not between MODIS and the HSI. The difference was between the HSI being reprojected when I loaded the MODIS data and the native UTM HSI projection. I would like to use the MODIS pixel footprints as an ROI for computing statistics (e.g., min, max, mean, stdev) on the HSI data. The HSI is in UTM (Zone 19N). I was able to convert the MODIS file to UTM but I'm not convinced that this resolves my footprint issue. I would expect the MODIS pixel footprint to be a parallelogram because of the scan angle. When I convert MODIS to UTM, the pixels are reshaped into 500-m squares which I'm guessing are not representative of the true MODIS footprint that accounts for scan angle. If the square footprint is correct, it is easy enough to draw a vector for one MODIS pixel and convert the vector to a shape file for use as an ROI to compute statistics. Here is a screenshot of showing the outline of an ROI that matches the square 500-m MODIS pixel. However, I'd like to do this for several hundred pixels. Is there an easier (more automated) way to compute statistics based on the 500-m pixel boundaries? I appreciate any help or insights. E.