[bahari]

dear all,
I am going to process an area and show subsidence phenomena there. but I see some errors/noises in out put. It seems we have some noises in the area of subsidence and find some points with +cm in velocity field close to other points with correct sign (-cm). that is first problem.
second is: I find some error around plain in hills that seems it is atmospheric effect. because I know those area must have zero deformation but I see -cm values in velocity field map.
I will send you the map.
finally I think it is necessary to select best parameters in “site processing” especially in the part of “Multi Image InSAR processing”. I couldn’t find good document on selecting the parameters of these part.
Please see if I am right on the errors/noises, help me on the steps of “Multi Image InSAR processing” and if you think the problem is related to other parts please lead me how can I find it?
best regards
Nader

Attachments:

You must be logged in to view attached files.

[periz]

Hi Nader,

about the second question, here you can read something:
http://www.sarproz.com/groups/sarproz-group/forum/topic/generic-question-about-ps-preliminary-selection/
Also, check the references mentioned there in the manual.

About the first question, we should find some time to have a remote session for looking at your data and at all steps you applied…

[bahari]

Dear Periz,
thank you for reply. but I couldn’t use your lead. Really I think it is necessary to have a better manual for software. for example I used below parameters in APS processing. I don’t know why some of options doesn’t work e.g. UW in this page and at the end GO key for APS estimate doesn’t work.
Please help me with details of options in APS processing & Sparse processing
thank you
Nader

Attachments:

You must be logged in to view attached files.

[matusbakon]

Hi Nader,

I would suggest to be patient. As in other processing windows you have to be familiar with what you are doing. Any of us that started with SARproZ has been experimenting with it for months until obtaining some reliable results, as its not a magic box where you press buttons and get outputs (like any other InSAR related software). First of all you are trying to estimate APS with 6 images? APS is estimated upon connections of the points (arcs) that you have defined by thresholding on Spatial Coherence. By including all zero values under Graph Creation you are using Delaunay graph, i.e try to exchange those values for 0 90 0 0 and you’ll see other type of graph – Flowered Tree – with the PSs in centers that are connected to the surrounding points in radius 90 [pix]. If you’ll change all the values e.g. Min Nr of points (e.g. 5) found within Minimum radius (25 [pix]) and Maximum radius (100 [pix]) with Radius step (25 [pix]) you’ll get user defined graph. What are the advantages of such types of connections? Then within network created upon such rules you are estimating Linear Trend without Height and without External DEM as reference? Is there any particular reason for that? In the end there are different types of APS you can deal with – Inverted Parameters (I.P.), Inverted Residuals (I.R.), Triangular Inverted Residuals (Tri I.R.), etc. I am sure you’ll find some comment on each option included in this window on this Forum as you’ve been allowed to research through it and for the second time advised to use it, see for example this:
http://www.sarproz.com/groups/sarproz-group/forum/topic/aps-processing-parameters/

Good luck and keep going on!

Matus

[periz]

Dear Nader,

here my comments:
1. yes we know that a complete manual is needed and we are working at that. in the meanwhile, please take seriously the available documentation. even if limited, it’s containing a lot of information and it takes time to be digested.
2. together with the documentation, you need to patiently read all the messages that the sw gives you. you will discover e.g. why some functions were not executed.
3. I do not know your current knowledge and experience in multi temporal insar. If you are not an expert, you have to take your time to learn things. For learning, please follow the tutorials. If you do not have data comparable to the one used in the tutorials, we can provide you with some samples. But you cannot pretend to take a set of data like the one shown here and be able to produce some reliable results right away. About the number of images needed in a standard multi-temporal analysis you can read on most of papers in the subject. You do not have enough images, so, the way to process your data is not standard.
4. I already said that for learning you need a step by step approach (and the tutorial are very helpful on this). If on the contrary your only target is to solve your problem at hand, then ask someone directly to check your data and suggest you how to process it. It’s a specific exercise, and it’s not useful to carry it out trough this forum.

dp

[periz]

Nader,

as a kind of general comment, with 6 images you should use a multi-master approach, and use also unwrapped interferograms. However, there are many comments more for your case, and also many options that could be chosen to help you solving your case. So, again, I suggest you to take a direct contact with some of us and have a remote session to look at your data and get suggestions
cheers
dan

[Author]

Posts