This tutorial is there to give the workflow necessary to compute MMASTER DEMs from ASTER L1A data. The method is described in details in the following paper : Girod et al, 2017
Part of the algorithms used in MMASTER rely on the ALGLIB library. Since it is licensed under the GNU GPL license, it cannot be part of the standard MicMac release (under CeCILL-B license). For that reason, it is necessary to compile MicMac from source, getting the IncludeALGLIB branch of the Git repository.
You also need to have Gdal installed on your system.
Any AST_L1A scene can work here, you can download them from NASA's Earthdata. When chosing the file format, choose GeoTiff (NOT hdf).
Workflows bash file (for now only for Unix systems) are available here : https://github.com/luc-girod/MMASTER-workflows . You probably want to add the folder where you have those files to your path.
Processing a single scene
Preparing the data
Examples are given using a scene named AST_L1A_00302212004225011_20170204145459_24750 and the UTM4N cartographic projection.
First extract the GeoTiff L1A to a folder named SceneName/RawData (ex AST_L1A_00302212004225011_20170204145459_24750/RawData).
Processing the scene
Then from the folder where **SCENE_FOLDER** (ex AST_L1A_00302212004225011_20170204145459_24750/) is situated run :
WorkFlowMMASTER_onescene.sh -s AST_L1A_00302212004225011_20170204145459_24750 -z "4 +north"
- extra options are possible with this file (given here with default values) :
-t 30 -f 1 -n false -c 0.7 -q 5
The RunMicMacAster_batch.sh" bash file can run MMASTER with the same UTM zone parameter for a folder full of zipped ASTER_L1A scenes.
Notes on default parameters
Default parameters are optimized for areas of low contrast (glaciers, sandy deserts...). If you are trying to get more details in your data, the correlation window size should be decreased (default "-q 5" means a 11*11 window). This can be done with the "-q" option, we found that "-q 2" gives the best results on well textured surfaces.