Loading...
 
ESA > Join & Share > TAO
Print

TAO

Introduction and Objectives

Current and forthcoming Earth Observation (EO) missions are steadily increasing volume, degree of variety and complexity, interconnection, integration and delivery rate of data. This poses challenges addressing the entire EO data lifecycle, from data collection, storage, management, dissemination and access to their analysis, exploitation, processing, integration and delivery to final consumers, leading to unprecedented opportunity to support and empower new types of user applications, information extraction and data analysis. The users as well are expanding widely their range of interests across multiple missions: optical, radar, multi-spectral, atmospheric.
A variety of toolboxes and libraries for remote sensing image processing are usually available to EO users to support and facilitate their processing needs. Furthermore, new requirements from EO data user communities are emerging, and in particular to be able to easily integrate their own manipulation, analysis and processing tools into an harmonised framework, which on the contrary should provide basic processing functionalities.

In order to build a consistent and light software framework which shall define and demonstrate the basis for next generation of EO toolboxes, the project shall:

  • Perform a survey on existing software tools/toolboxes, libraries and processing frameworks for EO images manipulation, analysis and visualiszation;
  • Perform a survey on existing user communities in the field of (optical) EO image processing and analysis, taking into account the widest possible spectrum of application domains and highlighting what are their needs in terms of functionalities;
  • Identify relevant standards and widely adopted interfaces, either Human Machine Interfaces (HMI) or Machine-to-Machine (M2M) Interfaces, permitting users to interact with and exploit functionalities of the software tools identified above;
  • Develop and implement a harmonised framework for integration and orchestration of existing and forthcoming processing modules, and the automation and paralleliszation of processing chains;
  • Demonstrate the effectiveness of the developed processing framework in relevant EO scenarios and use cases.

Expected Results

The TAO software framework shall allow extension and reusability of commonly used toolboxes (SNAP, OTB, etc.), permitting processing composition and distribution in such a way that end users could define by themselves processing workflows and easily integrate additional processing components.
TAO platform provides a mean for orchestration of heterogeneous processing components and libraries in order to process scientific data. This can be achieved in following steps:

  • Preparation of resources like the execution node, the processing components and data input;
  • Definition of a workflow as a processing chain;
  • Execution of workflow;
  • Retrieval / visualization of the results.

The TAO framework allows the following deployment modes:

  • Standalone mode where only a set of core components are deployed allowing execution of orchestrated workflows from an external toolbox that does not possess orchestration. The standalone mode can be assimilated also with the service mode of the TAO platform. In this mode, all the executions will take place on the local machine where TAO is installed.
  • Platform mode where the full platform is deployed, allowing users management, resource and catalogue and distributed execution on multiple nodes. In this mode, there will exist administrator users that will perform the configuration of the nodes and will set add accounts or set quotas for the users. The regular users will have the possibility to create, manage and execute their own workflows and also to visualize the data produced.

TAO frameworks allow combining in workflows applications from some standard supported frameworks like OTB and SNAP but also user custom scripts. The execution of the workflows can be performed on the local machine (in the case of Standalone mode) or distributed on several machines that were defined as execution nodes in the TAO (in the case of Platform mode).

Schedule

MilestoneDate Place Description
KO2016-07-01ESRINKO
M12017-02-14ESRINSRR
M22017-06-22ESRINPDR
M32017-12-07ESRINMTR
M42018-05-22ESRINCDR
M5 ESRINAR
M6 ESRINFP

Contractors

CS ROMANIA as Prime Contractor
CS SI as a subcontractor
Brockmann Consult Gmbh as a subcontractor

Deliverables


Contributors to this page: Michele Iapaolo .

Page last modified on Monday 04 of June 2018 16:08:38 CEST by Michele Iapaolo.