X
PrevPrev Go to previous topic
NextNext Go to next topic
Last Post 23 Aug 2016 01:48 AM by  anon
Processing of ALOS PALSAR for DEM extraction
 4 Replies
Sort:
You are not authorized to post a reply.
Author Messages

anon



New Member


Posts:27
New Member


--
23 Aug 2016 01:48 AM
    Greetings! I am interested to extract an INSAR DEM from 2 scenes of ALOS PALSAR. So far, I have changed the grid size to 10 m. I have a couple of questions in regards to the GCP selection step: 1) How many GCP is needed? Are there any tips other than using the .fint and .cc files? 2) Which refinement method should be used? Polynomial or automatic? Mine defaults in polynomial mode. 3) How can I validate the extracted DEM? Thank you and have a good day Regards Ben

    Deleted User



    Basic Member


    Posts:228
    Basic Member


    --
    23 Aug 2016 11:10 AM
    Hi Ben, 1. You will want to read the GCP Manual for SARscape that is provided by the SARscape developers at their web site: http://sarmap.ch/tutorials/GCP_Manual.pdf. The section on Refinement GCPs discusses the criteria for selecting good refinement GCPs, and many other details and suggestions to consider. The developers recommend that you use as many good GCPs as you can, with 20 or 30 being a good number if possible. The minimum number of points I believe is 7. But you will get much better results if you choose more points. 2. Polynomial is always the default refinement method. The SARscape documentation explains the different options like this: Refinement Method Different ways to perform the phase refinement (both in case of Digital Elevation Model and Displacement Map generation) are possible. One of the following methods can be selected to be used in the "Refinement and Re-flattening" processing step: - Automatic: by setting this flag the orbit configuration is first estimated on the basis of the input Ground Control Points. If the "A-priori check>Achievable RMS" is larger than the threshold, or the absolute normal baseline is smaller than the "A-priori check>Minimum Baseline", or the "A-posteriori check>Final RMS" is larger than the threshold, or the "A-posteriori check>RMS Ratio" is larger than the threshold, or the number of GCP is lower than 7, than the program automatically switch to the "Residual Phase" method; otherwise the "Orbital" correction method is applied. - Orbital: by setting this flag the orbital correction parameters are estimated using the Ground Control Points. Selecting this flag both the "A-priori" and the "A-posteriori" checks are disabled, while the only necessary condition for this method to work it is that at least 7 GCPs are available. - Polynomial Refinement: by setting this flag a phase ramp is estimated from the unwrapped phase without considering the orbit configuration. In this case the minimum number of Ground Control Points has to be equal to the "Residual Phase Poly Degree", otherwise the poly degree is automatically decreased accordingly. If you do not think of a reason to choose one of the non-default options, based on these descriptions, then it is usually reasonable to use the Polynomial option. 3. To validate the extracted DEM, you would need to compare it to an independently created DEM that has accuracy you trust. SARscape and ENVI don't have tools designed specifically to validate continuous images like DEMs. - Peg

    Deleted User



    New Member


    Posts:27
    New Member


    --
    23 Aug 2016 07:42 PM
    Hi Peg Thank you for clarifying and I note the point regarding more GCPs yielding better results. In this sense, what does 'better results' mean? Perhaps better accuracies ? Regards Ben

    Deleted User



    New Member


    Posts:27
    New Member


    --
    24 Aug 2016 01:10 AM
    Hi Peg I see, thank you for clarifying regarding GCPs. After processing the GCPs, there is an output window stating the mean and standard deviation errors. What range of errors here is acceptable? Also I have one other question regarding Sentinel 1 data import. For the precise orbit files, I have downloaded from the sentinel1 site in .EOF format. During import, do I simply select the corresponding .EOF file? The reason for my asking, is that I have selected the .EOF file, but I am unsure if it is even being used in the import process. There does not appear to be an indication for the precise orbit file being used. I look forward to hearing your advice. Have a pleasant day ahead! Regards Ben

    Deleted User



    Basic Member


    Posts:228
    Basic Member


    --
    24 Aug 2016 09:48 AM
    Yes, a more accurate estimation of the orbital correction that is needed.
    You are not authorized to post a reply.