Remote Sensing project...ideas

**bagtagley** · 03-30-2004, 10:58 AM

I've gotta come up w/ an idea for a project in my remote sensing class. I had a plan sort of, but it didn't pan out cause I couldn't find the data I needed. Basically what I need to do is use Erdas to manipulate/analyze a set of aerial or satellite photos. I was thinking about trying to do a change detection analysis on Mt. Hood to compare snow coverage 15-20 years ago vs. more recently, but I can't find photos that I don't have to pay for.

Can anybody offer up some ideas? More specifically, does anybody have or have access to data that I can use in a change detection analysis, image normalization, etc? Any help would be greatly appreciated.

**iceman** · 03-30-2004, 11:04 AM

Endless had a thread, I think it was back on powmag, where he was comparing snow cover on hood over various years, he should still have those pics, they're probably exactly what you need.

edit: either that, or I don't understand what you need, which is quite possible.

**bagtagley** · 03-30-2004, 11:13 AM

I'll take a look, thanks.

**iceman** · 03-30-2004, 11:17 AM

They might have been lost in the purge, but I bet he still has them, probably pm him.

**Telenater** · 03-30-2004, 11:29 AM

Originally posted by bagtagley
Can anybody offer up some ideas? More specifically, does anybody have or have access to data that I can use in a change detection analysis, image normalization, etc? Any help would be greatly appreciated.

Yes, I have access to data (or at least can point you to some free stuff). I have reasonably good access to data for California, an MIGHT be able to help out with other areas. Lets all pray that GWB doesn't get his way and move Landsat back to a "Cost of Acquisition" basis from it's present "Cost of Reproduction" pay scale. Regan tried that, and pretty much killed the US's remote sensing programs.

For CA landsat data check... here
It's one of the better Landsat data collections. Using remote sensing anywhere in the continental US to detect longterm changes in snowpack will be a significant challenge as snowfall has been stochastistic enough that a strong trend has not really emerged within the available data record. That, and the backscatter from snow makes remote sensing more problematic than you'd think. Also, be aware that if you try to compare imagery from winter to summer the shadows complicate things.

**bagtagley** · 03-30-2004, 11:38 AM

Thanks Nater

My hope was to find anniversary dates around 15 years apart. I was just going to try to do a band combination that would highligh areas of bare ground where there had once been snow.

I'm not at all set on this, it was just an idea. You got any suggestions for something a little more conventional?

**Telenater** · 03-30-2004, 11:51 AM

Originally posted by bagtagley

I'm not at all set on this, it was just an idea. You got any suggestions for something a little more conventional?

To get a longer time span you'll probably have to use TM data from LS 5 or normalize between TM and ETM+. I don't have any good sources for SPOT, or IRIS data which are the only other systems that have been active for that type of scale.

An easier (and a process I'm more familair with) would be vegetation change. High intensity changes such as forestfires and logging will be the easiest to accomplish with high accuracies. Another advantage of this method is that you'll be able to use data from the same sensor and have a higher chance of having training sites for normalization (large parking lots are your friend).

If you're still interested in snow RS you'll be doing a lot of reading about Dr. Jeff Dozier out of UCSB (also happens to be the guy I want to do graduate work with).

If you do choose to do your work in CA, I have a few older LS scenes that might be usable to create your time scale, but I don't know of any internet accessible sites for getting older landsat data for free.

If you have any more questions feel free to drop me a line at {screenname}@hotmail.com

**bagtagley** · 03-30-2004, 12:00 PM

I've got a stack a foot high of Dozier articles. Good luck, seems like he'd be good guy to work under.

I'll shoot you an e.

**Innominatus** · 03-30-2004, 12:00 PM

I'm trying to get a sense for what you're trying to do, but you're too far away.

My poor grasp of remote sensing aside, would it be easier to analyze something that has more readily available data and where the terrain itself changes, such as the 30-yr displacement of San Andreas fault or something?

**ak_powder_monkey** · 03-30-2004, 05:37 PM

Try contacting mt. hood I'm sure thye have that kind of data and could help you, As far as remote sensing goes I don't know if its any remote sensing maybe look at weather satalites because there is a ton of data from those...

I read today actully that there are satilite radars that can get snowpack measurements it might be cool to look at that if it actully exists, not just in therory.

**Buster Highmen** · 03-30-2004, 05:50 PM

Can you map Kimmie's boobs?

**Telenater** · 03-30-2004, 05:54 PM

Absolutely! Using the geology department's 3d visualization equipment and the Remote Sensing lab next door's hyperspectral spectroscope I can create a very precise model of the subject that is radiometrically correct and accurate to within a smaller tolerance than I'll be able to identify by hand (accuracy assesment and all that....).

**Buster Highmen** · 03-30-2004, 05:56 PM

Originally posted by Telenater
Absolutely! Using the gynecology department's 3d visualization equipment and the Remote Sensing lab next door's hyperboobstral boobstralscope I can create a very precise model of the subject that is boobtrometrically correct and accurate to within a smaller tolerance than I'll be able to identify by hand (accuracy assesment and all that....).

**bagtagley** · 03-30-2004, 05:57 PM

That'd be a good way to prove that reduction is a bad idea.

**midget** · 03-31-2004, 03:03 AM

I took a remote sensing class in college cause i had to. I hated it. I took a GIS class in college cause I had to.....I ended up being the TA cause I was the only one in my class that had a clue how to use a computer worth a shit. I hated it. I'm a geek, but not that much of a geek!

The USFS has to have that data, come on!

**jdabasin** · 03-31-2004, 08:11 AM

Originally posted by ak_powder_monkey
I read today actully that there are satilite radars that can get snowpack measurements it might be cool to look at that if it actully exists, not just in therory.

uhhhh....what language is this?

**jdabasin** · 03-31-2004, 08:17 AM

Originally posted by bagtagley
I've gotta come up w/ an idea for a project in my remote sensing class. I had a plan sort of, but it didn't pan out cause I couldn't find the data I needed. Basically what I need to do is use Erdas to manipulate/analyze a set of aerial or satellite photos. I was thinking about trying to do a change detection analysis on Mt. Hood to compare snow coverage 15-20 years ago vs. more recently, but I can't find photos that I don't have to pay for.

Can anybody offer up some ideas? More specifically, does anybody have or have access to data that I can use in a change detection analysis, image normalization, etc? Any help would be greatly appreciated.

Do a study on soil moisture changes over the last decade...or atleast the last few years in the western U.S.
It'd be interesting to see if we can qualify Denver as a desert area yet since we've had no rain for 2+ years it seems! It'd be interesting to see soil moisture changes from say 1980 to present to kinda get an idea as to how the climate has fluctuated, atleast regionally, since then.
Basically...we need some freaking precipitation before CO burns down this summer and maybe you can give us some meteorological karma!!!

**Endlessseason** · 04-02-2004, 10:40 AM

Originally posted by iceman
Endless had a thread, I think it was back on powmag, where he was comparing snow cover on hood over various years...

Mt Hood glacier comparisons, 60 years apart