X
PrevPrev Go to previous topic
NextNext Go to next topic
Last Post 01 Feb 2012 04:53 PM by  anon
SVM on Polarimetric Data
 1 Replies
Sort:
You are not authorized to post a reply.
Author Messages

anon



New Member


Posts:
New Member


--
01 Feb 2012 04:53 PM
    I'm working on Polarimetric RADARSAT-2 data and planning to classify them using SVM method. I have noticed that ENVI will take so much time to process my data even more than 3 days, to get the classification result. My area is 5000*5000 pixel with a spatial resolution of 8meter. I have 15 class and I have 3-4 region for each class in my training data. I was wondering if there is a solution for this problem to run SVM-ENVI faster on this data set.

    MariM



    Veteran Member


    Posts:2396
    Veteran Member


    --
    02 Feb 2012 07:57 AM
    SVM is a very processing intensive algorithm and is known to take a long time to process larger datasets with many classes. If you have many ROIs and those ROIs contain a large number of pixels, then the process takes a long time for the convergence. You can try to reduce the number of pixels in your training ROIs to improve performance.
    You are not authorized to post a reply.

    Forums Products ENVI