FLAASH PROCESSING OF LANDSAT TM - NV5 Geospatial

Prev Go to previous topic

Go to next topic

Last Post 29 Jun 2010 10:20 AM by anon

FLAASH PROCESSING OF LANDSAT TM

4 Replies

Sort:


You are not authorized to post a reply.

Author

Messages

anon

New Member

Posts:

29 Jun 2010 10:20 AM
HI, I am very new to this forum, I am trying atmospherically correct Landsat TM using the FLAASH Module of ENVI. I ahve been getting some problem and need your suggestions 1. what should be a radiance scale factor, and whether it should be ally to all bands in one time or it should be implemented band by band. But I use this as 10 because I read it in some of posts in the forum. 2. In FLAASH parameters, I am confuse about Ground Elevation(km) what does that mean and I can not find any information about this parameter in the header file provided by usgs. In FLAASH processing of Landsat TM tutorial it was taken to be 0.1, so I take it as 0.1 3. In advance settings, also the zenith angle is automatically picked by the module which is 180 degrees and also about this one too, no information is available in header file. 4. also in I read scene center as (CENTER 1055829.7581E 100652.1262N), when I put these parameters a prompt up windows pop up with message (calculated solar elevation for the given date, time, lat,long is negative. with a value of approximately -77.90, note longitude should be negative for western hemisphere). to me I myself feel there may be something wrong with scene center lat, long, as they looks very strange to me because the longitude can not not exceed 90 degrees as far as I am concern. please help me where i am getting wrong, I will be greatly thankful to you all in advance,

Deleted User

New Member

Posts:

29 Jun 2010 10:53 AM
Hi Cheny, FLAASH is a rather complex atmospheric correction tool which means that you need to know and understand the data you are putting into it as well as understand what scale factors are for and how to apply them. Throw in any of the advanced settings and things can get quite messy. I usually try to make things as easy as possible when doing atmospheric corrections. Why complicate things if you don't need to right? My thoughts... Given that you have Landsat data it may be easier if you use the QUAC module in ENVI. Everything is set up for you so you don't have to worry about scene center and all that. Just so you know you probablby need to convert your scene center from eastings and northings to actual Lat and Lon values. Ciao

Deleted User

New Member

Posts:

03 Jul 2010 05:00 PM
Hi, thanks for your previous reply, I did try your suggestion by using the QUAC to correct my images atmospherically. My task is actually, to apply three indeces (NDVI, NDWI and NDSI) to extract different landuse patterens. the problem is that when I apply NDVI, the value for the NDVI which I get is very strange (Min: -5.000000 and Max: 6.750000) while it should be between -1 and 1. the same thing happend with NDWI whose values I got (Min: -10.200000 and Max: 4.027027) while it should also be between -1 and 1. I will be greatly thankful to you if you can explain men where I am getting wrong Regards Cheny

Deleted User

New Member

Posts:

13 Aug 2010 09:05 AM
Hi, To answer your questions: 1. Use 10. That is the scaling factor by ENVI after Landsat data calibration to radiance. It works. 2. I use Google Earth to find the approximate ground elevation of my area of interest in the scene. Ideally we'd be able to use a DEM for this, but FLAASH just isn't there yet. Choose the elevation which is the most representative of your target. In some cases the elevations wary between 0.4 and 3 km, but what I care about is in the 0.5 km range, hence that is what I use. 3. Don't mess with the advanced settings unless you absolutely know what you're doing or have been instructed to do so by someone in the know. You only really need to mess with these for pointable (e.g., Quickbird) or hyperspectral (e.g., Hyperion) sensors, and even then, only when necessary. 4. Make sure all coordinates are in latitude and longitude. The image center coordinates can be calculated from the midpoints between the upper left and lower right scene corners. I get these from the MTL file, which is a text file. The same goes for first line time. 5. I would not use QUAC if you can use FLAASH, as it assumes too much and doesn't let you configure anything beyond sensor type. Best regards, Guy Serbin

Deleted User

New Member

Posts:

13 Aug 2010 09:10 AM
I have one question/ comment: Are you running QUAC on NDVI/ NDWI imagery? You should only run atmospheric corrections on multispectral data and not index data. Those indices should only be calculated from atmospherically corrected data.

You are not authorized to post a reply.