X
PrevPrev Go to previous topic
NextNext Go to next topic
Last Post 29 Jun 2015 03:47 PM by  anon
MLC using spectral libraries.
 1 Replies
Sort:
You are not authorized to post a reply.
Author Messages

anon



New Member


Posts:
New Member


--
29 Jun 2015 03:47 PM
    Hello. I am trying to run a maximum likelihood classification (MLC) using a spectral signature that was taken using a field handheld spectrometer. I have input the spectral signature into a resampled spectral library and everything looks good. However, I do not know how to use the spectral library as input to my supervised classification. Is there a way to run an MLC with a spectral library or signature? This seems possible from my understanding. MLC uses ROIs to calculate an average spectral signature and then classifies the remaining pixels based on those signatures, right? So is it possible to bypass the ROI and imagery based spectral library using an external spectral signature? Chad.

    MariM



    Veteran Member


    Posts:2396
    Veteran Member


    --
    30 Jun 2015 12:37 PM
    MLC doesn't just use an average, it uses covariance and probabilities which must be calculated from multiple samples in a training data set (or for an ROI, a group of pixels). So MLC cannot use a single mean spectrum and there is no way to use a 'group' of spectra as a class in ENVI. However, the minimum distance and SAM classifiers do not have this requirement and can use library spectra to classify.
    You are not authorized to post a reply.