A new Open Source Toolbox for Interferometric Time-Series Processing and Data Analysis
Marinkovic, Petar1; Perski, Zbigniew2
1PPO.labs, NETHERLANDS; 2Polish-Geological Institute - National research Institute, Carpathian Branch, POLAND
The "market" of open-source processors for interferometric time series analysis is fragmented. Most of the solutions are internally developed by some research institute, and while, most of them being open-sourced, and occasionally even extensively documented, they lack a refined and approachable design, as well as algorithmic flexibility. They are usually designed and built on top of the, state-of-the-art, but heavily specialized algorithms, and having specific applications in mind. Therefore, any further algorithmic development or extension, development and deployment additional features, or setting-up of a processing facility on top of these solutions is very difficult.
Since beginning of 2013, Polish Geological Institute - National Research (PGI-NRI) Institute, runs a pilot project for monitoring a terrain surface deformation of gas exploration sites in Poland by means of the interferometric time series analysis. During the project preparation, and setting up of an algorithmic and processing setup, we identified aforementioned limitations. All these, led us to recognize an opportunity for this and other internally funded projects, as well as the whole community. We decided to continue with setting out a higher ambition, to create a modular and open-source toolkit, for InSAR time-series processing and analysis. Our goal is to provide a refined, well documented, and extensive library of flexible algorithmic and design components for internal development. Furthermore, our objective is to make it available to the community (under permissive open-source license), to build and innovate on.
Specifically, we identified building blocks of the publicly available algorithms (eg., SBAS, StaMPS, DePSI, etc.), and abstracted them into components of the library and application interface (API). This library is then used to construct, in a "lego-building-blocks-like" manner, processing flow that is executed on top of the input data. All modules share the same data and memory model, for each of them an extensive unit-test cases are available, and all of them are extensively documented. The share data model makes the all the components reusable and exchangeable, unit tests allow iterative development and extensive testing, while the documentation (coupled with unit-tests) provides an explanation how to operate, use the library and toolkit, and build and innovate on top of it.
As the basis for this development, jDoris library is used to aggregate and unify all the algorithmic deployment into a single API, jDoris processing engine is used for distributed processing, while NEST (Next ESA SAR Toolbox) is used for the visualization and analysis of the results. Note that jDoris is currently being utilized for all InSAR functionality of NEST.
The initial release, of this new toolbox, is scheduled for an early summer of 2013, that will provide a set of basic tools for processing and analysis of interferometric stacks for deformation monitoring applications.
Apart from presenting the development philosophy of the toolbox, this contribution will in detail review the available and planned functionality, development challenges, and community building efforts.