In this lab, we will be working with a variety of Landsat TM and Radarsat images. Mostly just visual analysis. Play around with band combinations, zooms, etc.
Start by downloading and unzipping the data (J:\geog430\fun.zip). Make sure you have lots of space - it will take close to 100 MB to download and unzip this data. Thus, work of the D drive to make your life easier.
1) Let's start with the radarsat (1 band) data. I have very little metadata. Ok - no metadata. The area is part of Kalimantan (Indonesia). You will note the rivers show up extremely well - water absorbs radar data extremely well. In general, you're seeing topography. Give me a paragraph or two describing what you see. Tell me about "layover" and why it exists(zoom in on the hills and observe - they seem to be sagging to one side).
2) Next are the two Roebuck Images (for more info, see my mudflats website). Basically, there are two dates of imagery (we're up in northern Western Australia). You have only TM bands 3, 4, and 5 (as 1,2,3 in your subset). Give me a couple more paragraphs: what sort of cultural features do you see? physical? changes between the two dates? general differences? etc... In other words, what do you see?
3) Now onto some coastal stuff - Cape Bougainville, northeastern Western Australia. Think equatorial and wet.You have TM bands 1-4. What do you see? Describe the landscape. What do you think is that large whitish area in the NW? A couple more paragraphs, please.
4) Finally, the Hamersley area in Western Australia. Very dry. Lots of iron mining. You have TM bands 1-5 and 7. In this case (unlike the last lab), the geology and landforms are very distinct. Zoom in on the mountains - you can see bedding in the rocks (especially in which bands?). Write a couple more paragraphs about what you see in the physical landscape.
Next, a little compare and contrast. On the data drive, geog430\ikonos\ subdirectory, there are a bunch of files:
They are pretty large, esp the 1m data - you might want to copy all this to the D:\ drive and work with it from there.
The question - compare and contrast the three images - what can you see, what not, why (remember your 4 types of resolution!)? Look at 'urban', agriculture, and range areas. In the Ikonos data, what do you think that white stuff over the small lakes/streams to the north is?
OK - on to GIS/RS integration.. Run unsupervised classifications on the following files: tm_no_dem.img and tm_with_dem.img. 20 classes (J:\geog430\unsuper_with_dem\). BE SURE TO WORK OFF THE D:\ DRIVE - THE FILES ARE HUGE! Give me a paragraph or two telling me the differences in the two classified images. Has the addition of a DEM to the image helped? Hindered? Explain all this in detail referring to specific landcover types (shrub-steppe, forest, irrigated ag, basalt, etc).
Due Wednesday, 31 May. 5 points.
Take your time, look at these images using different bands (where available), zoom in/out, etc and give me detailed writeups.