X
PrevPrev Go to previous topic
NextNext Go to next topic
Last Post 08 Jun 2012 08:27 AM by  anon
Constructing a confusion matrix on an already classified image
 1 Replies
Sort:
You are not authorized to post a reply.
Author Messages

anon



New Member


Posts:
New Member


--
08 Jun 2012 08:27 AM
    Hello, I produced a classified image using a decision tree through R. The output from R was a .envi file that I converted to a .dat. I would like to use ENVI's confusion matrix tool to evaluate the results of the classification. I have .evf files to use for ground truth ROIs. However, ENVI tells me I must use an ENVI classification in order to construct a confusion matrix. Is there a way to remedy this? Should I somehow reclassify the classified image in ENVI first? Not sure how to do this while preserving the decision tree results...

    MariM



    Veteran Member


    Posts:2396
    Veteran Member


    --
    08 Jun 2012 12:50 PM
    If the image is a classified image but ENVI does not recogize it as one, you can change it to an ENVI classification image by going to File ->Edit ENVI Header ->Edit Attributes ->Classification Info. You will be prompted to specify the number of classes. If you have 8 real classes, set it to 9 so that there is an 'unclassified' class which is always required. Then you will be presented with a dialog that allows you to specify names/colors for the classes. Save the header and now it should be treated as an 'ENVI Classification' image.
    You are not authorized to post a reply.