sarproz-group

PSInSAR Analysis

This topic contains 13 replies, has 2 voices, and was last updated by  anarizka 2 years, 2 months ago.

  • Author
    Posts
  • #894

    anarizka
    Participant

    Dear Prof.Daniele

    I make scatter plot in SARPROZ, and select interested area in “Small Area Processing” tool. I have some problem to understand:
    1. Point that high coherence value must in building area, but in this result, the Google Earth Image and Point don’t match exactly. Based on your tutorial, your GE image very clear, but from my result GE image very rough. Could you explain to me why the point don’t match in GE and how to get GE image clearly?

    2. In Geocoding window, I try to export data in csv.format. If I make replica (ex.2), I get different results on the same date (100 data), but If replica 0, I only have 20 data. I use 20 ASAR image. What this meaning of replica and its advantages?

    Please, find attach file below.

    Best Regards,

    Ana

    Attachments:
    You must be logged in to view attached files.
  • #898

    periz
    Keymaster

    1. to have the points match in GE, you need 2 things
    – first of all, you must have (previously) chosen a GCP that aligns your SAR data to external geographic features (as a DEM or as points selected in GE)
    – secondly, if you are processing a small area, the reference point should have been previously processed in the “site processing”, and its height must have been correctly estimated. if this is not the case, your data will look like shifted
    – if the shift happens because of that (the wrong height of the reference), you can manually re-align your data in the small area by manually inputing a height correction in the box Href (geocoding module)
    – please note that from your attachment the small area does not look like much small, and points decorrelate far from the reference point
    2. if your phase is phi, a replica is phi+2pi. sarproz takes 1 replica up and one down. so, 1 replica corresponds to phi-2pi and phi+2pi. 2 replicas (your choice) means exporting phi-4pi, phi-2pi, phi, phi+2pi, phi+4pi which makes 20*5=100 data

  • #902

    anarizka
    Participant

    Dear Prof.Daniele

    To have point match in GE, I have done first step to choose GCP that align in SAR image. you can see in 1.jpg (attach file). I think its SAR image and GE match well. After I choose small area processing, the shifting problem still unsolved (2.jpg file). As far as processing, I still can not find where is the Href box located.

    1. Could you tell me where location it is?
    2. How we can know how large/height correction must input in the box Href (geocoding module) manually?

    Best Regards,

    Ana

    Attachments:
    You must be logged in to view attached files.
  • #905

    periz
    Keymaster

    Hi Ana,

    from your geocoded refl map 11.jpg, I can tell you that it is not well geocoded. It looks shifted.
    The Href is at the top of your 21.jpg. Currently, you see NaN inside the text box.
    To know the value of the height, you need to sketch the SAR acquisition geometry, considering the incidence angle.
    Do it, it’s a good exercise for understanding how it works…

  • #906

    anarizka
    Participant

    Dear Prof.Daniele

    I have done to choose Href, prof.
    From the first step about geocoded refl map, I must re-process it again or no. Because from GCP selection, the SAR coordinat and Geo coordinat in GE it looks like match well.

    Please, find attach file below

    Best Regards,

    Ana

    Attachments:
    You must be logged in to view attached files.
  • #911

    anarizka
    Participant

    Dear Prof.Daniele

    1. For make sure that PSInSAR correct processing, how many factor must be considered besides from the attach image?
    And how about your opinion based on my results
    2. From PSInSAR, Must I have GPS data for verification? Any likelihood below:
    (a). I have GPS data but the statiun far away from my area (equal to 8km)
    (b). I did not consider about GPS data for PSInSAR analysis.

    Best Regards,

    Ana

    Attachments:
    You must be logged in to view attached files.
  • #915

    periz
    Keymaster

    Hi Ana,

    your result is not looking very good. How big is the area you are analyzing? can you plot the PSC graph with coherence-colored connections? Can you show us the reflectivity map and the coherence map?
    dp

  • #916

    periz
    Keymaster

    Hi Ana,

    your result is not looking very good. How big is the area you are analyzing? can you plot the PSC graph with coherence-colored connections? Can you show us the reflectivity map and the coherence map?
    dp

  • #917

    anarizka
    Participant

    Dear Prof.Daniele

    I have processed almost 3 square-kilometer area.
    Hope you give me suggestion to increase good results.
    How about my second question (GPS data). Could you explain it?

    Please, Find attach file below.

    Best Regards,

    Ana

    Attachments:
    You must be logged in to view attached files.
  • #926

    anarizka
    Participant

    Dear Prof.Daniele

    Another question.
    I try to compare with GPS statiun (LIKNstatiun that I plotted in GE). This statiun have distance around 8km from my area. I think its shape almost same. So, Have I considered it or no?

    Best Regards,

    Ana

    Attachments:
    You must be logged in to view attached files.
  • #932

    periz
    Keymaster

    Hi Ana.
    the selection of points looks not good. so, the graph has few good connections and the result is quite bad.
    if you are analyzing a 3×3 sqkm area and you have a GPS station at 8km distance you cannot do much.
    dp

  • #939

    anarizka
    Participant

    Dear Prof.Daniele

    I have tried using another parameter for “Images Graph=Delaunay Graph”.
    In APS processing I used DSxDL is 0x0 and matr.Coher.Win is 10×10 also used I.P. (Inverted Parameters) options. Hope you give me suggestion for results.

    Any other question
    After I made scatter plot, I was rather confused about analysis relation of amplitude and temporal baseline. Could you tell me how far it was very contributed in PSInSAR?

    Please, find attach file below.

    Best Regards,

    Ana

    Attachments:
    You must be logged in to view attached files.
  • #948

    periz
    Keymaster

    this result looks much more reasonable.
    however, I’d suggest you to use the I.R. APS. read old posts on this topic.

    I do not understand your question about the amplitude time series.
    Your example looks ok and it shows that no evident pattern is visible.
    For other cases, please read some papers about temporary PSs (or check the ampl time series change detection tutorials)

  • #980

    anarizka
    Participant

    Dear Prof Daniele,

    I choose second area to analysis. Based on my result, have you think that it reasonable or no? Please, find attach file below.

    In APS estimation, Any parameter unknown, could you explain what is “Estimated Parameters like, R r0 ds, Flatten, R and DSF” also in “Processing Parameters like Scattering Centers and Polynomial Order” ?

    Best Regards,

    Ana

    Attachments:
    You must be logged in to view attached files.

You must be logged in to reply to this topic.