Grab the kit_8_00 image from the data drive - you will be using this for most of the lab. It's Landsat TM imagery collected in August of 2000. Note - this and the derivitive images will chew up a lot of space. I'd suggest that you work from the D drive.
Alternatively, you could set up a smaller AOI and work from that on a zip or something.
FILTERS
Mash the 'Interpreter' button at the top of the page, then 'Convolution'. In this popup you will see a gajillion different filter options - you can even make your own. If you want to see the exact kernel for the filter, hit the edit button (you will want to do this later). Note - often the results make more sense if you draw one band at a time (ie - band 1 in red, green, AND blue).
Run the 3x3 low pass, edge detect, horizontal edge, vertical edge, and Laplacian filters (all 3x3). Look at the results. Play around a bit.
Q.1 What is the kernel (how it treats the 3x3 'hood) for each filter? Compare and contrast the results of the filters. When might the output be useful?
RATIOS
Again, mash the 'Interpreter' button, then 'Spectral Enhancement', then 'Indices'. This gets you in the 'play with ratios' page.
On the kitco image, run the NDVI, Iron Oxide, and Clay Mineral ratios.
Q. 2 Take a look at the output (the output is only 1 band). What do they appear to be mapping (as opposed to what they should be mapping - IF there's a difference)? Give me a paragraph or two discussing the output.
Next - we play with a little Aussie imagery. In this part, we will be working with a Landsat TM dataset collected over north-western Western Australia (the Kimberley). It is a very dry region.
Start by grabbing the data (data drive, 430 subdirectory, kimberly.zip) and unzipping the data.
Examine the area in, first, a false colour image (3,2,1). Then as a colour IR (4,3,2) image. This is just to get a bit of familiarity with the area.
Q.3) Vegetation is shown quite clearly in the colour IR image. Explain this difference. When would this knowledge be useful to a geologist or geomorphologist?
------------------------------------------------------------------------------------------------------------
Redraw the image using bands 7 (listed as band 6), 5, and 1.
Q.4) Compare and contrast the false colour and 751 images. Where is 751
a superior image for geologic interpretation. What about the colour IR image?
------------------------------------------------------------------------------------------------------------
Now, let's look at a few ratio images. To generate these, you will hit the interpreter
button, then spectral enhancement, then indices. Run the following:
Be sure to check the 'stretch to unsigned 8 bit' box.
Q.3) what features should each image highlight (as a result of the ratio)? In other words, given what you know about what the different bands tend to be good at identifying (high or low R), what do you think these ratios are designed to highlight? However, what is it that they seem to be highlighting on these particular images?
6 points. Due on Monday, 24 April.
Worth 6 points. Due wednesday, 9 February.