sarproz-group

APS Processing problem

Viewing 2 reply threads
  • Author
    Posts
    • #6325
      haonan
      Participant

        Hi everyone,
        I still have a question about APS processing. This time I use TSX(2013-2015, 32 images). I remember that: ‘the value of the coherence in the temporal plot is useful only to look at the trend in time, to check whether there are some left signals that were not well modeled.’ I think this time should be this situation. So how should I change the parameters to make connections processing result looks better?

        Thank you so much!

        Attachments:
        You must be logged in to view attached files.
      • #6327
        haonan
        Participant

          By the way, I also want to ask under these situation, how do we choose Non-Linear Weighting parameters. I usually use the default parameters(0.71 0.84 0.97).

          Attachments:
          You must be logged in to view attached files.
        • #6333
          periz
          Keymaster

            1. the coherence trend in time becomes visible when you plot the coherence as a function of time. this plot is generated when you display the final test, after APS estimation and removal. the first plot that you show is a histogram, there is no time variable there
            2. the histogram you show is not bad. it shows 2 peaks: signal and noise. now you have to keep the signal and discard the noise
            3. non linear weights are exactly the way to discard noise and keep signals. try the default numbers (which are calculated based on your current set of data/parameters) and go on with the next processing steps. if APS is ok and tests are ok, you are done.
            best

        Viewing 2 reply threads
        • You must be logged in to reply to this topic.