Grab a copy of Chapter 5 (unsupervisecd classifications). Work through it.... Remember to work of your personal drive or the D drive.
Next, we start to play with some Landsat data.
Grab the kit_8_00.img from the geog430 subdirectory on the data drive. It's a TM image of this area. You will need over 100mb free on your disk for this lab - IF you're careful about file management. Keep an eye on this.
Fire up the imageinfo window - the little page icon with the 'i' on it
What's the spatial resolution?
For each band: what are the minimum and maximum reflectance values. Do you see multiple peaks? If so, what do you think is causing it? (to assess this, view just that band (same band in R, G, and B) - this will give a greyscale view). (Mash the 'I' button to get all this info)
Play around a little - zoom in/out, change the bands, etc.
Next, do a unsupervised classification. Run it three times - 10, 15, and 20 classes
Which of these three runs seems to best represent what is on the ground? Why? What do you think are the different classes? (pick the run that seems the best to you and list the classes and what you think they are). I will be comparing your list to your unsupervised classification. Print out, in color, both a copy of your final supervised classification and the associated table (with new colors and class names). (or make a legend and use that.
Worth 6 points. due monday, 1 May.