X
PrevPrev Go to previous topic
NextNext Go to next topic
Last Post 21 Oct 2008 10:43 AM by  anon
ENVI Feature Extraction Forum
 3 Replies
Sort:
You are not authorized to post a reply.
Author Messages

anon



New Member


Posts:
New Member


--
21 Oct 2008 10:43 AM
    Post your questions, comments or ideas about ENVI Feature Extraction here.

    Deleted User



    New Member


    Posts:
    New Member


    --
    19 Mar 2010 11:52 AM
    How do I process a large file.  I tried help and searching this forum, but I can't figure this out.  Is there a way to process chunks of the file sequentially instead of the entire image?  Thanks, Dan File: Clipped a 2009 USDA raster .sid image of Montgomery Co, Ohio down to a 391M file of the City of Dayton. It is now in .img format.  Hardware: HP  xw8600 quad-core workstation with 16M of Ram.  Software, ENVI 4.7, ENVI EX, ENVI Zoom. Process: ENVI 4.7 Zoom -> Feature Extraction workflow. 

    Deleted User



    New Member


    Posts:
    New Member


    --
    06 May 2010 09:17 AM
      I could not answer this question.  I think that there is also too much variability across the image for Feature Extraction to process accurately.  One can make out the suture-like borders of the adjacent collect paths in this mosaic image.   I have now used an existing set of shape files for the City of Dayton and run ArcMaps/ArcCatalogue, Spatial Analyst extension, Extraction, Extract by Mask (ie shapefile) [Tool only works if one navigates to it from ArcCatalogue, not ArcMap!!!]  I subsequently ran it again against each neighborhood.  

    Deleted User



    New Member


    Posts:
    New Member


    --
    06 May 2010 09:32 AM
    Normal 0 false false false EN-US X-NONE X-NONE MicrosoftInternetExplorer4 Feature Extraction and Mask I now have smaller (.img) files with which to work.  These images are comprised of an irregularly shaped neighborhood surrounded by a broad border of black pixels.  I assume that this is a condition of raster formats being a rectangle of row and column pixels. Feature Extraction now runs in a reasonable amount of time, but it fails to produce any usable results.  Mostly it produces a single feature class for 99% of the image.  When I run Feature Extraction on a ‘chip’ of the data, then I get good polygons.  The chip has no border of black pixels.  Do I need a Mask to ignore these black pixels for Feature Extraction?  I have not been able to quite follow the Help file for Creating Masks.  Should I continue spending time on this strategy?  Is there any trick for pulling all of those black pixels out as the mask?
    You are not authorized to post a reply.