X
PrevPrev Go to previous topic
NextNext Go to next topic
Last Post 14 Nov 2008 03:44 AM by  anon
Felling detection using alos palsar data
 3 Replies
Sort:
You are not authorized to post a reply.
Author Messages

anon



New Member


Posts:
New Member


--
14 Nov 2008 03:44 AM
    Hey my first post here and a newbie to the envi enviroment also... What i want to do is try to compare "alos palsar images" (plain images), the images are separated by time and sometimes they dont overlap perfectly. The comparison should be on the difference between db before cutting and after cutting... via a simple thresholding algorithm. This would preferably generate a vector (or raster) outlining the edges of what can be identified as clear cuts. This would then be compared against actual polygons drawn from SPOT and aerial photographs used as reference when identifying clear cuts. Im a little bit stuck in how to proceed after getting the images properly georeferenced and generating an image that contains data in decibels and also making an overlay of the vectorlayer...? Any help would be appreciated and of course quouted in the thesis im working on.

    MariM



    Veteran Member


    Posts:2396
    Veteran Member


    --
    14 Nov 2008 11:11 AM
    I don't know what to tell you about what thresholds to use, but if you have an idea of the thresholds to use, then perhaps a first attempt would be to make a density slice of these images.  You can do this via the Display menu ->Tools->Color Mapping ->Density slice.  Here you can set various thresholds based on the image data values.  So areas between 0-10db might be one range, and 10-20db the next and so on.  This applies a single color to those pixels that meet the threshold (like a classification).  These ranges can be exported to a classification image which can then be exported to vectors.  Would sokmething like this work?

    Deleted User



    New Member


    Posts:
    New Member


    --
    17 Nov 2008 06:31 AM
    Thanks for the reply! Density slice seems to be a step in the right direction! now its just the problem of finding the right thresholds but that should be pretty straightforwards. This will probably work out fine! Now its just the question of processing alot of images i guess... are there scripts or addons avaliable for running batches or something similar? Also it should be enough using one header and replicating and renaming it for all of the images right?

    MariM



    Veteran Member


    Posts:2396
    Veteran Member


    --
    17 Nov 2008 08:38 AM
    All batch processing in ENVI is handled through IDL and ENVI doit routines.  Have a look in the ENVI programmers guide for more assistance.  If all your images are the same, you can replicate the .hdr for all input files.  This is assuming the files are in ENVI format.
    You are not authorized to post a reply.