Loading...
 
ESA > Join & Share > Technology Projects > UGEI Project
Print

UGEI Project

Project Title   Universal Geometry Engine for EO Images
Project Acronym   UGEI
Contractor(s)   INDRA Espacio, Institut de Geom├ática

 

Project       Context
         Objectives
           Architecture
         How it works
       Output
 

  


Context

Within EO context one of the main issues to be solved prior to the effective use of products imaging the Earth surface is to locate them with reference to the scene being observed. Another common need of EO products is to geometrically relate them with each other.

  • Application projects where EO imagery is used usually imply multi-source data integration: images and cartography (raster-vector), or fusion of images (raster-raster) are just examples. 
  • Correction of geometric distortions are outstanding for these kind of activities to allow for element comparison, registration and overlay. The impact of distortions are becoming a more important problem with very high resolution satellites.

From the geometric point of view, both issues mentioned above can be stated in the same manner: how to relate or transform known source coordinates to unknown destination coordinates. This relation can be represented as a geometric sensor model.

In such a transformation several matters are involved: knowledge of the platform position at every instant during the image acquisition, of the particular sensor configuration, of the type of scene being imaged and the particular scene, of the state of the atmosphere at that moment. 

The Universal Geometry Engine is a system able to perform the estimation (in a least squares adjustment sense) of a geometric model for a given set of entries, where the model is a function that relates source coordinates with destination coordinates (image coordinates with terrain coordinates), and the entries are a specific configuration of the imaging system or systems and a set of reference data.

Top  


Objectives

UGEI project objective has been the development and testing of a Universal Geometry Engine whose aim is to provide to the EO user with the means needed to ensure a coherent geometry among a set of images, each one acquired by a given sensor, so that they can be jointly processed.

UGEI project was preceded by preparatory phase called UGEIP whose main objective was the preliminary design of UGEI system.

The objectives of UGEI system, from functional point of view, are to:

  • Extract information from auxiliary satellite image product files on:
    • Exterior orientation (orbit, attitude)
    • Internal orientation (sensor geometric configuration)
  • Provide the user with the capability to generate observation data: Ground Control Points and Tie Points
  • Perform the estimation of a geometric model parameters based on previously identified input data
  • Provide the user with exploitation application of adjusted model parameters: orthoimage production through a generic orthoimage engine,

Main and innnovative design aspects of the system are the combination of the following features:

  • The extension of the geodetic/photogrammetric network methodology and its underlying mathematical theory to remote sensing
  • The use of a sensor geometric rigorous model representing the physical reality of the image acquisition geometry. These models account for the effects of the platform (position, velocity and attitude), the sensor (viewing angles), the Earth shape and the relief, so that the influence of each of the parameters can be physically interpreted. Sensor models of different types are implemented in UGEI: ERS SLC and SPOT-5
  • Extensibility: incorporation of new sensor models or observation types do not impact on system architecture. Sensor models are encapsulated in cartridges with an standardized interface.
  • Genericity: each and every model are internally treated in the same way,
  • Deployment
    • desktop
    • selected components are prepared to be deployed in network environment
    • processing components provides an API facilitating its use from an external application

 

Benefits expected upon achievement of functional and design objectives are:

  • Reduction of dependence on external tools 
  • User knowledge and control of the modeling (capability to act on model parameters)
  • Identification of some repetitive patterns in sensor modeling
  • An universal geometry modelling approach with standardised modeling concepts

Top  


Architecture

The UGEI main components are the following

 

  • Ingestion of orbit, attitude and sensor calibration data from product image data
  • Observation generation: ground control points and tie points.
  • Numerical core able to perform optimal estimation of model parameters.

 

The output from UGEI are used by the Orthoimage Generation application: based on the estimated model parameters and a DTM, produces an orthoimage in a cartographic system selected by the user.

UGEI system architecture has a 4-tier architecture as the one shown in the following figure:

Image

 

The role of each layer is the following:

  • The upper layer, control layer, implements the mechanism for the execution of algorithms developed in the computation layer.
  • The GUI layer implements the interface to the user.
  • Computation layer offers an API to the upper layer and encapsulates the mathematical algorithms.
  • Libraries layer contains a set of shared libraries used by upper layer components.

Top  


How it works

The figure below depicts the main processes and data-flow within UGEI . Also included is an application of UGEI output: the orthoimage generation application.

Image

Sequence of operations sketched in the diagram can be summarized as follows:

  1. Product image of a vendor of a adequate level is ingested in the system, that means:
    1. Conversion of image data to tiff format
    2. Extraction of orbit, attitude (EOP) and sensor calibration (IOP) from product auxiliary image data and transformation to internal UGEI format (.ing)
  2. Operator using an appropriate GUI measures GCP and TP over images, those data are stored for ulterior processing (.gmf and .imf files)
  3. The steps 1 and 2 can repeated as many times as required by operator
  4. Using block building application, operator is able to select the information acquired in steps 1 and 2 to generate the set of files input to Network adjustment: instrument (.inst), parameters (.par) and observation (.obs) files. Using this data Network adjustment numerical core performs optimal estimation, in the least square sense, of sensor model parameters for a set of selected images
  5. Using adjusted parameters computed previously and a DTM an orthoimage is generated

     

Top  


Output

Samples orthoimages generated by UGEI (left: SPOT5, rigth: ERS 1)

Image

The software was installed at ESRIN and presented at a final presentation on 15th October, 2009.


Contributors to this page: .

Page last modified on Thursday 16 of December 2010 09:02:49 CET by .