Indra Espacio

1
Indra Espacio
( Área reservada a imgen )

• UGEIP
• M3, Frascati ESRIN
• 15-October-2004
• Final Presentation

2
Agenda

• Introduction
• Technical objectives
• UGEIP work description
• Architecture of implementation
• UGEI work description
• UGEI testing and validation

3

• Introduction

4
Introduction General Information
UGEIP INDRA proposal 1047/03-IE generated in
response to RFQ/3-10814/03/I-OL
named UNIVERSAL GEOMETRY ENGINE
PREPARATION UGEIP
5
UGEIP PROJECT OBJECTIVES (as per KOM)

• Analysis of UGEI system context to explore and
  analyse the potential geometric capabilities of
  system.
• To perform a trade-off based on Analysis of UGEI
  context information in order to identify the
  system capabilities.
• Specification and design of a UGEI system and its
  components
• generic/universal numerical engine core that
  performs optimal least-square sensor model
  parameter estimation based on a set of input
  data sensor type, orbit, attitude and ground
  control points.
• collect input data to feed numerical engine core
• exploit the numerical engine outputs
• Identification and definition of sensor models to
  be used by engine core and/or exploitation
  applications.

6
UGEIP DESIGN OBJECTIVES (as per KOM)

• Main
• Extensibility system to design will be flexible
  enough to incorporate new sensor models without
  modification of architecture.
• Multisensor capability allowing handling of
  several sensor models simultaneously
• Secondary to be analyzed in the frame of the
  project, likely to be considered for
  implementation, in order of precedence (later
  discussed in the frame of presentation)
• Deployment web/grid, stand-alone (Main TBC)
• Portability

7
Introduction Companies involved

• Prime Contractor
• INDRA ESPACIO (Spain)
• External services consultancy
• Instituto de Geomática

8
Introduction Project Team in INDRA organization
9
Introduction Project Team
10
Introduction Project Phases

• First phase
• UGEIP (current one)
• Exploratory and preparatory nature
• Intended to define the Requirements and
  Architecture for a Universal Geometric Engine
  (UGEI).
• Start 29-12-2003
• End 15-10-2004
• Second Phase
• UGEI
• Development of UGEI
• Planned start 31-12-2004 (TBC)
• Estimated duration 18 month (To be Consolidated)

11
TECHNICAL OBJECTIVES OF UGEI

• UGEI Universal Geometry Engine
• Contents of this part of the presentation
• Review of the geometric modelling problem
• Conventional solutions
• Motivation
• UGEI solution
• Characteristics, advantages and benefits of the
  UGEI solution

12
PROBLEM DIMENSION

• Geometric modelling Problem dimension

13
PROBLEM DIMENSION GEOMETRIC MODELLING I

• GEOMETRIC MODELLINGRelate EO imagery to known
  reference frames
• Mathematically relate 2D image space and 3D
  object space
• Sensor specific problem

14
PROBLEM DIMENSION GEOMETRIC MODELLING

• Some image space to object space transformations
• 2D / 3D image referencing
• Orthorectification / DEM generation
• Stereo- / Mono-plotting
• Some object space to image space transformations
• Computation of image coordinates for an object
  for image analysis
• Projection of given cartography to image space
  for updating
• Interferogram simulation for differential SAR
• Sensor simulations

15
PROBLEM DIMENSION SENSOR MODELS I

• Modelling conceptsAnalogue and digital cameras
• Well established concepts
• Photogrammetric standard
• Physical models
• Modelling strategy
• Collinearity model
• Self calibration

16
PROBLEM DIMENSION SENSOR MODELS II

• Modelling conceptsAcross track or whiskbroom
  scanners
• Sensor dependent
• Landsat series
• AVHRR, GOES,
• Physical and empirical models
• Modelling strategy
• Depending on spatial resolution grade
• e.g. physical models supported by2D / 3D
  polynomial functions

17
PROBLEM DIMENSION SENSOR MODELS III

• Modelling ConceptsAlong-track or pushbroom
  scanner
• Sensor dependent
• SPOT series
• IRS series
• HRS
• Physical and empirical models
• Benefited from photogrammetry
• Modelling strategy
• Physical modelling
• HR refinement with polynomial functions or 3D
  rational functions

18
PROBLEM DIMENSION SENSOR MODELS IV

• Modelling ConceptsSAR sensors
• Key role in information exploitation
• geo-coding
• radargrammetry
• Interferometry
• Physical models
• Modelling strategy
• Rigorous modelling

19
PROBLEM DIMENSION CURRENT SOLUTIONS

• Current Solutions
• Multipurpose SW packages (EO remote sensing )
• Application specific solutions (photogrammetry )
• Sensor specific solutions (line scanners, SAR
  ...)
• Current Limitations
• No general, extensible tool for geometric
  modelling problems
• No support of flexible / individual development
  and adaptation of sensor models

20
PROBLEM DIMENSION NEW REQUREMENTS

• New requirements to geometric modelling
• High precision requirement
• Need for physical models
• Sensor / data fusion
• Growing number of applications
• Growing number of sensors and platforms
• New missions

21
THE UGEI SYSTEM

• Approach to UGEI

22
MOTIVATION FOR UGEI

• Purpose of UGEI
• Provide an universal SW tool for geometric
  modelling of imagery to support information
  exploitation, value adding and innovation in PRS
• Issues
• Extensibility
• Standardization
• Unique treatment of geometric modelling problems
• Portability

23
UGEI SYSTEM INNOVATION

• Extensible with
• Observations and control information input
  channels
• Observation generation tools
• Sensor Models
• System application modules (information
  exploitation tools)
• Standardized in
• Network adjustment and sensor modelling
• Reference frame and coordinate system
  transformation
• Network object
• Portable to
• External information exploitation tools
• Laboratory, workshop and factory environments

24
UGEI SYSTEM
25
UGEI SYSTEM
Extensibility
26
UGEI SYSTEM
Extensibility
Standardization
27
UGEI SYSTEM
Extensibility Unified Modelling
Standardization
28
UGEI SYSTEM
StandardizationPortability
Extensibility Unified Modelling
29
CHARACTERISTICS OF THE UGEI SYSTEM

• Characteristics of the UGEI system

30
THE NETWORK

• Network
• Set of observations, parameters (unknowns) and
  instruments, inter-related by mathematical models
• Network adjustment (network solution)
• Optimal estimation of network parameters and
  their covariance
• Method
• Iterative non-linear least-square estimation on
  the general model
• Unique treatment of all geometric problems

31
THE SENSOR MODELS

• Extensible set of sensor models
• General model, extensible and adaptable
• Basic set of sensors
• Classical photogrammetric model
• SPOT-5
• ERS-PRI
• Characteristics
• Part of the network
• ? Usable in exploitation applications
• ? Image-to-object and object-to-image
  transformations

32
THE INTERFACES

• Network adjustment configuration through GUI
• Configuration of computation through user ?
  configuration of computation and solution
  strategy ? configuration of gross error
  detection ? facilitates simulations
• Network visualization and manipulation
• Browsing and editing of networks ? advanced
  network control and easy modifications ?
  facilitates sensor model development and
  optimization

33
THE OBSERVATIONS

• Observation generation tools
• Built in general, light image matching tool
  (GIM-lite)
• GIM-lite characteristics
• GCP and tie-point collection for optical sensors
• Graphical user interface
• User initial identification of Tie Points.
• Approximation of homologous location

34
THE APPLICATIONS

• Information exploitation
• Built in ortho-image generator
• Extensible with system application modules (SAM)
• DEM generation
• Stereo- / Monoplotting
• 2D image georeferencing
• 3D image georeferencing
• others

35
OTHER FUNDAMENTALS

• Built on solid geodetic, physical and mathematic
  grounds
• Reference frame transformations
• Coordinate system transformations
• Based on international standards
• ISO
• EUROSDR standardization efforts
• Platform independent deployment
• Isolation of deployment dependent code in special
  libraries

36
BENEFITS OF UGEI SYSTEM

• Benefits of the UGEI system

37
BENEFITS OF UGEI

• An advanced system for geometric modelling for
  all types of experts in PRS
• The scientist in the PRS laboratory
• Has a tool to generate new models, missions and
  applications
• The technician in the PRS workshop
• Has a tool to generate new workflows or modify
  procedures in data processing
• The operator in the PRS factory
• Has a tool for geometric calibration and
  orientation of P RS data

38

• UGEIP Work Description

39
UGEIP TASKS (I)

• Task 1
• Management of project PMP generation
• Generation of web pages for promotion
• Document configuration and quality assurance
• Task 2
• Analysis the context where UGEI is intended to
  work in order to determine the potential
  candidate functions for the system
• Task 3
• Generation the UGEI SW requirements document
  (SRD) for the system, justifying the choices for
  the functional scope justification of UGEI.
• Task 4
• Generation of Architecture design document (ADD)
  with contents recommended by ECSS standard for
  Design Definition File (DDF).
• Generation of ADD technical note justification
  for the adopted solutions with contents
  recommended by ECSS for Design Justification File
  (DJF).

40
UGEIP TASK (II)

• Task 5
• Preparation of Verification and Validation Plan
  Methods and Tools document with clear
  identification of input data needed for UGEI test
  execution.
• Elaboration of plans for UGEI implementation
  phase
• Task 6 SW Management Plan with provisions for
  development, testing and configuration management
  of the SW.
• Task 7 Technical implementation proposal
  describing the activities to be performed for the
  development, integration and validation of the
  system, the development and testing strategies,
  the needed COTS if any.

41
UGEIP Project Logic
42
UGEIP Baseline Schedule
43
UGEIP Updated Schedule
44
UGEIP MEETINGS
 
 
 
45
UGEIP PROJECT DELIVERABLES
 
Legend D Draft U Update F Final version  
46
Summary of deliveries(I)
47
Summary of deliveries(II)
48

• Architecture

49
SYSTEM OVERVIEW

• UGEI
• Ingestion of orbit, attitude and sensor
  calibration data from product image auxiliary
  data (EO sensors)
• ? PRODUCT INGESTION
• Collection of ground control points (GCP) and
  matching of tie points (TP) on images
• ? OBSERVATIONS GENERATION
• Numerical core perform optimal estimation of
  model parameters based on observations and sensor
  information
• ? NETWORK ADJUSTMENT
• Ability to incorporate new modules in the system
• ? CONFIGURATION
• Add on UGEI application ? ORTHOIMAGE GENERATION

50
SYSTEM CONTEXT
51
SYSTEM DESIGN (I)
52
SYSTEM DESIGN (II)

• 4-tier architecture
• 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.

53
INTERFACES

• Standalone system not directly interfaced to any
  system.
• UGEI subsystems loosely coupled interchanges
  between subsystems based on files.
• Input files to be used by a S/S are selected via
  GUI.
• Internal files used within UGEI will be XML
  formatted
• Types of files internally managed 
• Observations files
• Parameters files
• Instruments files
• Configuration files
• Images TIFF
•  

54
UGEI Coding Languages

• Java
• S/S GUI
• Control layer
• C
• Computation layer
• Library layer
• Some COTS used coded in C (PROJ4)

55
UGEI COTS
 
56
UGEI Development Tools
 

• IDE

• UML design tool TBD
• Java GUI development tool TBD 
• Configuration Control CVS

57

• UGEI Work Description

58
UGEI Task Definition

• Design consolidation
• Acceptance Test Plan
• Coding and Unit Testing
• UGEI Testing
• INDRA factory system and system Integration
  Testing
• Acceptance Test Cases and producedures
  Preparation
• Test Execution
• User/developer manual Preparation Missing in
  PMP draft
• Project Manag./Quality Ass./Conf. Control

59
UGEI Project Logic
60
UGEI Tasks Details on Design Consolidation

• Sensor Models Mathematical definition generation
  of STBD (software theoretical basis document)
• UGEI ICD
• UGEI DDD
• Consolidation of ADD document based on inputs
  from previously generated documents to generate a
  draft version of DDD for CDR with enlightened
  contents.
• Generation of an As built DDD for M7
  (acceptance review) using code documentation.
• DDD elaboration Use of Javadoc/Doxygen to
  generate subordinates section of each major
  component/subsystem/application of ADD
• Approach for generation of DDD in line with is in
  line with applicable ECSS standards to this
  project.

61
UGEI Schedule (I)
62
UGEI Schedule (II)
Notes Estimated duration 18 month (TBC by
proposal lt 24 months). Proposed plan too tight.
63
Meetings Reviews
 
Note that first progress meeting is numbered as
M4 (last milestone of UGEIP was M3)
64
Meetings Progress Meetings
 
Frequency of progress meetings is bi-monthly when
no intermediate review is between two consecutive
ones. Some progress meetings as a teleconference
for the sake of cost and travel
optimisation. First progress meeting is numbered
as PM3 (last meeting of UGEIP was PM2)
65
Deliverables Documents

2 ESA to confirm the need of generating this
document.
66
Deliverables SW and Data Items
 
Open point HW to run UGEI?
2 ESA to confirm the need of generating this
document.
67
UGEI Implementation Risks (I)

• Planning delay
• Proposed planning too tight
• Each WP needs different skills/profiles of
  personnel
• Maximum use of free SW COTS tools is proposed.
  Risks in the development if some errors are
  detected on these tools. The probability of such
  a risk is reduced if it is considered that
• Tiff libraries are used by most of image
  processing packages
• Xerces tool for XML has been used by INDRA
  without problems in another projects
• Log4j/Log4C is a well-tested and very extended
  tool both for Java and C applications world.
• PROJ4 is an extended coordinate transformation
  package and maintained.

68
UGEI Implementation Risks (II) Planning
Delay/Cost

• UGEI desktop deployment in Windows (or LINUX)
  environment -gt Reduction of testing effort
• Renounce to code (budget, time) some auxiliar
  components. Removal do not affect UGEI main
  objectives. i.e
• Configuration file editor
• Task manager allow for launching several
  processes simultaneously
• Delivery of UGEI to ESA
• SW delivery
• HW delivery (TBC)

69
UGEI Open Points

• Intended use at ESRIN
• Factory Testing
• Witnesssed by ESA
• Modifications ?
• Acceptance testing at ESRIN, as a demo.
• Post-Delivery issues.

70

• Testing and Validation

71
UGEI Testing

• SVVP Methods and Tools defining
• Unit test scope
• Testing strategy
• Testing method
• Test sequence
• Test data set
• SVVP to be completed during UGEI implementation
  phase

72
UGEI Testing

• Unit testing
• Integration testing aimed at verifying the
  correct
• Build up of S/S starting from components
• Build up of System starting from S/S
• Factory System testing of the integrated
  components (witnessed by ESA)
• On site Acceptance testing (ESRIN)

73
UGEI Unit Testing

• Black-Box Testing
• Software units of Libraries layer
• Dynamic loader of configuration S/S
• Task manager of control S/S

74
UGEI S/S Integration Testing

• Black-Box Bottom Up Strategy
• Integration of S/S components shall be done in
  two steps
• Library integration- Computation layer
  integration (yelllow box)
• Integration of previous build with graphical
  interface (red box).

75
UGEI System And Acceptance Testing

• Black-Box testing
• Acceptance Testing Subset of System Test

76
UGEI Components not directly tested COTS

• Coordinate transformation package PROJ4
• Geotiff Library
• Xerces XML parser
• Logging packages log4j and log4c

77
UGEI Testing Documentation
78
El valor de la anticipación