FAQ

GENERAL QUESTIONS
Which radar satellite data does DIAPASON process?
What are the languages supported by DIAPASON?
My DIAPASON run has failed, what should I do?
How to create a multilook image?
Is it possible to process ASAR IM RAW data with PRISME?
Can DIAPASON generate interferograms with ALOS data combining double and simple polarisation?

DATA EXTRACTION / IMPORT
What are the sensors/formats supported by DIAPASON?
What are the format supported by DIAPASON?
How do I convert images from one format to another?

ORBIT RELATED QUESTIONS
What are the orbits state vectors supported by DIAPASON?
How to calculate the baseline of a dataset?

FOCUS RELATED QUESTIONS
How to focus ENVISAT Image mode data?
How to focus JERS data?

DEM RELATED QUESTIONS
In what case should I use a DEM: D-INSAR or INSAR?
How to supply a DEM as input?
Does DIAPASON have a specific module for phase unwrapping?

COREGISTRATION RELATED QUESTIONS
How to control the precise azimuth time and range distance obtained?
Master / Slave coregistration has failed, what should I do?

GEOCODING RELATED QUESTIONS
With a fake DEM, the interferogram has to be calculated in the SAR geometry (inter_sar) or can it be calculated in the DEM geomtery (interf_mnt)? When we use interf_sar, how do we geocode the image?
To have a geocoded interferogram, can I run interf_sar and then use the ortho tool to geocode the phase?

GENERAL QUESTIONS

Which radar satellite does DIAPASON process?
   The radar satellite data processed by DIAPASON Version 4.3 are:
   - Raw data: ERS, ENVISAT and JERS.
   - SLC data: ERS, ENVISAT, JERS, RADARSAT-1, ALOS and TerraSAR-X
   Radarsat-2 data processing will be available with next release.
   Contact us for more information on the release schedule: diapason@altamira-information.com.

What are the languages supported by DIAPASON?
The native language of DIAPASON is French. However since V4, the English language is fully supported.
The environment variable LANGUE determines the language to be used.
Set it the to:
- FRANCAIS for French
- ENGLISH for English
by using under bash (or sh) shell the following command: export LANGUE=ENGLISH.

My DIAPASON run has failed, what should I do?
The first thing to do, is to examine the log and error message and to determine the faulty processing step.
In many cases, the problem is related to of the issues described below.

How to create a mutlilook image?
There are at least 3 possible ways:
1 - The easiest way is to use the DIAPASON GUI. In the menu "Tools/Multilook from SLC"
2 - Using in command lines the script ml.sh provided in the template directories
3 - Using directly the utility cp_ima (copier) and applying the following OUTPUT transformation:
- in English: ave, quad, azi=10, ran=2; conv, ty=byt, gain=0.01
- in French : moy, quad, azi=10, dist=2; conv, ty=byt, gain=0.01
Please refer to the manual for further details.

Is it possible to process ASAR IM RAW data with PRISME?
Yes, since DIAPASON V4.1 (july 2006) a new extraction module has been added to decode ASAR Source Packets.
Please see the FAQ in the PRISME section. However, it is necessary to create a precise orbit prior to do the Doppler centroid estimation and focussing.

Can DIAPASON generate interferograms with ALOS data combining double and simple polarisation?
Yes, DIAPASON can generate interferograms with ALOS data combining double and simple polarisation under the following constrains:
- The combined images must have the same polarisation (HH or VV).
- The resolution of the images to be combined must be similar.

DATA EXTRACTION / IMPORT

What are the sensors/formats supported by DIAPASON?
DIAPASON supports RAW and SLC data for ERS, JERS and ASAR (Image Mode)
DIAPASON supports SLC data for RADARSAT-1, ALOS and TerraSAR-X sensors. These files should be imported into the DIAPASON format by the use of the routines EXTRACT and CONCAT.
EXTRACT creates the annotation (geosar) and orbit file.
CONCAT generates the data file on the DIAPASON format that is commonly CI2 for SLC data and C5B for RAW data.
JERS RAW data is coded in c3b.
The particular routine EXTRACT_RAW_ASAR is used instead for extracting and decoding ASAR RAW data.

What are the format supported by DIAPASON?
DIAPASON supports CEOS format for ERS, RSAT, ALOS, TerraSAR-X and JERS data.
DIAPASON supports Envisat format for ASAR and for ERS data.
For ERS CEOS formats some foreign stations provide CEOS format that is not fully compliant with the ESA guidelines. DIAPASON could have problems in extracting those files. For further details on ESA format please look at: http://earth.esa.int/services/esa_doc/doc_sar.html.

How do I convert images from one format to another?
DIAPASON image format is raster with an associated external descriptor file called RAD file. The utility CP_IMA (COPIER) allows to transform any DIAPASON format with the possibility to apply (or not) a transformation. The commonly used DIAPASON formats are:
- byt (or oct), i2, i4, r4, r8 and for complex date c3b, c5b, ci2, cr4.
Please refer to the manual for further details.

ORBIT RELATED QUESTIONS

What are the orbits state vectors supported by DIAPASON?
1- Orbit state vectors annotated in the original file and extracted by the routine EXTRACT.
2- Precise Delft orbits for ERS-1/2 and Envisat are supported. They should be translated in DIAPASON format by the means of the routines GETORB and DELFT2ORB.
3- For Envisat DORIS files can also be used as well. The DORIS2ORB utility should be used in this case.
Further information on DELFT orbits could be found at: http://www.deos.tudelft.nl/ers/precorbs/
Since DIAPASON V4.2, the GUI integrates a tool for automatically dowloading DELFT precise orbits from a GEOSAR.

How to calculate the baseline of a dataset?
The ALT_AMBIGUITY (or ALT_AMBIGUITE) utility allows computing the baseline components for all the possible combinations. Geosars are required as input.
Thus the extract step has to be made first.

FOCUS RELATED QUESTIONS

How to focus ENVISAT Image mode data?
1- Use the EXTRACT_RAW_ASAR to extract at once the GEOSAR and the data file. A template script is provided on this purpose
2- Download the precise orbit
3- Perform the Doppler parameter with DOPPLER_RATE
4- Use process with PRISME using the GUI or the prisme.sh template script

How to focus JERS data?
For JERS data, prisme requires 2 more arguments which are the I and Q channel bias. This compensation avoids a phase error in the processed image. Compute the mean I and Q on a block of RAW JERS data. The as JERS RAW data are coded in C3B format the Bias I and correction are simply obtained by doing:
3.5 - mean_I and 3.5-mean_Q.
On the geosar, the presumming factor has to be increased to a value >=3. This will reduce the size of the SLC image by the given ratio but will in the same time improve the phase quality.

DEM RELATED QUESTIONS

In what case should I use a DEM: D-INSAR or INSAR?
When you want to make D-INSAR for ground motion monitoring (subsidence, volcanoes) a DEM has to be given as input in order to remove the topographic phase component. On the contrary, if you want to make classical InSAR for DEM generation (as example), there is no need to supply a real DEM. However, a fake DEM is by the way necessary to remove the flat earth component phase. A fake DEM can be created by the routine create_dem (affect_mnt). Images hereafter shows interferograms of the same zone obtained with and without DEM.


D-INSAR interferogram	InSAR interferogram

How to supply a DEM as input?
You first have to create a DEM descriptor and then to update the fields: DEM TYPE and DEM DESCRIPTOR into the master image geosar. DEM TYPE has to be set to TRUE.
Or you can use the routine CREATE_DEM (AFFECT_MNT) and answer NO to the question "Creation of a fake DEM?"

Does DIAPASON have a specific module for phase unwrapping?

DIAPASON uses Snaphu for phase unwrapping, which is one of the best 2D unwrappers for interferometry. It is easily accessible via the GUI. For further information please consult http://www-star.stanford.edu/sar_group/snaphu/

COREGISTRATION RELATED QUESTIONS

How to control the precise azimuth time and range distance obtained?
The first 3 DIAPASON steps SIMU_SAR, CORREL_SIMU and CORRECT_STNR (CORMOY) are necessary to correct the master image from the inaccuracy of the original timing annotations.
After the extraction steps and till the CORRECT_STNR (CORMOY), the parameters of the geosar NEAR RANGE (NR) and TIME OF THE FIRST LINE have to be considered as rough ones and thus have to corrected.

After CORRECT_STNR (CORMOY) the STATUS of the master geosar is set to PRECISE and the fields PRECISE NEAR RANGE and PRECISE TIME are updated. These parameters are calculated by comparing the master multilook image with a master multillok image simulated with a DEM (SIMU_SAR). The shifts measured by correlation (CORREL_SIMU) are afterwards transformed in slant range distance and azimuth time correction times (CORRECT_STNR).

The precise time stamps can be controlled by performing the DEM lock test. This test consists in recreating a new simulated image by putting SIMU_SAR as start/stop step. The simulated image will be now made using the precise time annotations. Then visually controlling that new simu image is perfectly overlapping with the real master multilook.

If it is not the case control the following parameters on the log files:
- at CORREL_SIMU step: control the shifts automatically computed and eventually supply them manually
- at CORRECT_STNR (CORMOY) step: control the number of control points kept. Usually a number of points > 1000 is OK. If not, decrease the threshold on the DIAPASON.sh or on the GUI.

Master / Slave coregistration has failed, what should I do?
You should look in the log file for more information on the failure:
- At CORREL_IMA step: control visually the shifts given by the automatic computation. If they are wrong then enter them manually on the DIAPASON.sh file by setting ROUGH_LOCKING_SAR at NO and supplying good shifts.
- At GEO_GRID (GRILLE_GEO) step: control the number of valid points. This number should be > 800 for a full scene. Typical values between 1000-2500 are obtained. The number of 'grip' points can also be increased by decreasing the threshold. However, this will have also the effect of adding more noisy points in the coregistration.
In the case of very similar slave images, another solution could be also to decrease the multilook factor of the ML images till the limit of 1x1. Reducing the multilook factors means also increasing the processing time.

back to top

GEOCODING RELATED QUESTIONS

With a fake DEM, the interferogram has to be calculated in the SAR geometry (interf_sar) or can it be calculated in the DEM geometry (inter-mnt)? When we interf_sar, how do we geocode the image?
Doing interf_dem (interf_mnt) with a flat DEM geocodes the image but does not ortho-rectify it. Meaning that the geomtrical distortions as foreshortening and layover are not corrected.
For an interferogram doing interf_mnt (or similarly for a SAR image doing calmnt (ortho)) with a flat DEM is almost the same as having an ERS GEC or ASAR IMG product.
In the version 4.2 DEM module, a new ortho program has been developed that takes as input an interferometric height (from inter_sar) and then geocodes it in the proper way in order to have an image geocoded and ortho-rectified.

To have a geocoded interferogram, can I run interf_sar and the use the ortho tool to geocode the phase?

The phase ortho-rectification shall be made by interf_dem. Ortho does a non-coherent averaging which is not suitable for phase.

back to top