| 16 Sep 2010 07:20 AM |
|
|
Dear Users,
I'm sabber from Bangladesh having a problem in Target detection. Whenever i want to detect any mineral using spectral libray(same libray for all mineral) with different methods (ACE, CEM etc) every mineral shows always same area. Every and each time all minerals shows the same Region of Interest (ROI). But it is not possible. I think its related to Rule threshold of Step 7 of target detection wizard. Whenever i change the value , ROI also changes. Can anybody please help me how to determine the Value of Rule Threshold for Different Mineral for different menthods.
|
|
|
|
MariM
Veteran Member
Posts:2396  
| 16 Sep 2010 11:44 AM |
|
|
The rule threshold is just a slider that allows you to include/exclude the selected pixels. If all of the default results look the same, then perhaps your imagery is not sufficient to distinguish between the mineral spectra you are using. Are you using hyperspectral imagery? How was your spectral library created?
|
|
|
|
Deleted User
New Member
Posts: 
| 16 Sep 2010 12:08 PM |
|
|
Dear Kitty,
Thanks very much for your reply. I'm using ASTER SWIR images and USGS spectral library that is built in . I have 220 band Hyperion data but i cant use those only for not to resample spectral libray resampling. Because in Spectral resampling wizard Hyperion Sensor information is absent in case Perdefined filter function is selected.
|
|
|
|
MariM
Veteran Member
Posts:2396
| 16 Sep 2010 12:34 PM |
|
|
ASTER has been used successfully to map some minerals. The spectral library in ENVI is in reflectance between 0-1 (0-100%) and the units are in micrometers. You would also need to convert your ASTER imagery into reflectance if it is not already in reflectance. The ASTER_07 product is a surface reflectance product. This product needs a scale factor to be applied to it to convert the integers to floating point reflectance values between 0-1. ASTER 1B data is in radiance and would need an atmospheric correction applied to convert radiance to reflectance.
|
|
|
|
Deleted User
New Member
Posts:
| 19 Sep 2010 04:11 AM |
|
|
Dear Kitty
Still i have the same problem. I have converted my ASTER imagery into reflectance. At this time No of ROI reduced but same this all minerals occupy the almost same area only few 10 to 100 pixels different. Please help me to tell me how to do this. I have already downloded all tutorials related to it from ITTVIS and followed Target Detection Wizard instructions. But i could not do it . Please help to tell me the preprocessing of images that i need. I have done atmospheric correcing, Reflectance conversions.
|
|
|
|
MariM
Veteran Member
Posts:2396
| 20 Sep 2010 07:19 AM |
|
|
I tested the tool with some ASTER data in reflectance and it returned different ROI areas for a couple of minerals (alunite/kaolinite). You say you get 'almost' the same areas but almost does imply that it returns different ROI areas. It may mean that the minerals you are trying to find with ASTER are not distinguishable. I really don't know how to help you with this. You might contact your local technical support.
|
|
|
|