search

ppolabs / jlinda

jLinda - Java Library for Interferometric Data Analysis
9 stars 5 forks source link

Cross InSAR functionality #8

Closed pmar closed 10 years ago

pmar commented 11 years ago

Core class and functionality implemented and committed. Operator needs to be profiled and further tested.

pmar commented 11 years ago

80% work finished with: https://github.com/ppolabs/jlinda/commit/3bfe2dd276cb3613b66e37bc41896564bf1fd9cf . There is still functionality for updating of metadata to be merged from non-public branches, and issues with random crashing to be resolved.

Still, I will not continue building this operator on top of NEST/JAI functionality. Will come back to it after new coregistration is being released. I need new resampling for this.

pmar commented 10 years ago

Closed with: 3bfe2dd + 129f109

pmar commented 10 years ago

Operator UI still needs some work. CrossResampling has an Action missing (see disussion at https://github.com/ppolabs/jlinda/issues/21). This is developed in some of local branches - integrate and publish.

pmar commented 10 years ago

Major rework of Cross Intereferometry functionality

Ref. commit: 76a942d709319c5cb322e28e8f8cfd629246c900 .

Summary:

The fnc is now more generic - instead of working on the stack of coarse coregistered SLC's the operator is re-worked to resample an image from the input geometry into the other one, that is selected by the user. In this case from ERS/ASAR geometry to ASAR/ERS respectively.

The idea behind this rework is that the operator can be reused - by plug in of the other deterministic relationships between other SAR geometries - for other cases that are similar to resampling needed for optimal Cross-InSAR. For example Deskewing of 'Actual Doppler' acquired ALOS data.

Workflow example: cross_insar_workflow

Testing

Operator/fnc tested on South Holland ERS/Envisat pairs. Coherence is there, but very likely something is going wrong with orbit handling. See Examples.

Example 1: Result of Cross Interferogram using annotated orbits.

cross_insar_results_2

Example 2: Result of Cross Interferogram using DORIS orbits for ASAR and annotated for ERS2. Changes in the fringe pattern suggest that ERS2 orbits are not reliable for InSAR. Unfortunately, NEST cannot 'ApplyOrbit' for ERS2 using EIGEN-GRACE01S. Issue is communicated to @lveci - the main developer of NEST.

cross_insar_results_1

Known Issues: