SNAPHU Software

Hi everyone!

I am working on DEM generation with sentinel-1 data. However I want to know that how can I use snaphu in windows environment? because when I run snaphu in snap, it makes error. Could you help me please?


you need to install cygwin for it. We recently talked about it here:

I made a tutorial for my students for compiling SNAPHU in Windows a time ago. It was originally written in German so the screenshots may not always totally match but the description should be sufficient.

Installation_SNAPHU_English_ABraun.pdf (849.1 KB)

Please note that this is no longer needed because snaphu 1.4.2 now has an already compiled exe file
Please see here:


thanks for your helps.


Hi there, have been following your tutorial to compile SNAPHU - thanks for making it available. I was wondering if you could help me with one of the steps (2e). When I type make install it says “make: command not found”. I have done some Googling but can’t seem to find a solution. I installed Cygwin as advised in the tutorial.

make sure that C:\cygwin64\bin (or wherever you installed cygwin) is in your PATH variable.

make.exe should be available in this folder. If not, run the cygwin installation again and install the make environment (step 1h in the pdf).

Okay will do the above, thanks a lot for your help!

Dear Dr ABraun

Thank you for the tutorial snaphu in cygwin, it is very helpful for me. But I have problem when unwrapping tile is not complete (until column 3). Please refer to attach file.

Waiting for your help


31000 lines is quite much. Did snaphu run until the end? According to your screenshot the unwrapped img file (corresponding to hdr) was not yet created.

Maybe you can also adjust the number of tiles in the snaphu config file.

Thank you for your fast response Dr ABraun

should I make it small area of interest? how much is proportional lines in range could be use (in your tutorial 1600 lines).

At the last part of your tutorial (g) also mention, after completion, a new unwrapped phase.snaphu.img file was created. so I have to wait until unwrapp finish, please correct me if I am wrong

Thank you for your help

try to export a small subset (2500 lines) and see if it works correctly. If yes, increase the subset to the extent that is interesting to you.


It works, It takes times 6 hours for 31000 lines.

I did for 23726 and take time 5 hours.

and here is the result in .kmz file after TC. It is not create fringes at all.

Thank you for your comments

strong ramp in this one. If you look closer, is there any topographic information on this one?
Did you apply topographic phase removal? What is your goal?

I did topographic phase removal.

My goal is to calculate fringes based on interferogram achieved.
I want to see landslide movement in that part (Central of Aceh District). Based on ground data. there are a lot of landslide in October 2015.

did you have good coherence between the images? Maybe you can post a screenshot of the coherence raster and its histogram.

Which snaphu method did you allpy? TOPO, DEFO or SMOOTH?

You could also consider making additional interferograms based on other image pairs, before and afterh October, just to get an impression on the general information which can be retrieved from the data.

1 Like

I attached of histogram in Takengon city (Near Lut Tawar Lake).

I apply DEFO (Deformation in Snaphu method)

I have 5 data:


I sellect 21 Sept 2015 as a master image, and 2 Dec 2015 as a slave image.

I have not repeat interferogram to other slave image yet.

I studied that large land movement for 2 m above can destroy fringes. but I am still lack study about atmospheric artefacts can be destroy fringes too.

theoretically, only the white areas in your coherence image reveal usable information. Then there are still atmospheric disturbances left which could affect your fringes. Try to apply a coherence mask on your interferogram as described below to see if the areas of interest are within that areas.

If the areas of interest are excluded by this mask, you need better coherence. You can achieve that by using an image pair with smaller temporal baseline (6/12 days are currently the smallest for Sentinel-1). You can also try to apply a better co-registration in order to improve the image alignment.

1 Like

I will work on that.

There only 24 days available data S1A on that area. I could not have 12 days or 6 days data

So how to do better coregistration.

Thank you for answering all my questions

maybe one of these points helps:


24 days is simply too long to allow for higher coherence. 12-day data should be available over your area by now.

1 Like