International Charter Products To KML

From OTBWiki
Jump to: navigation, search

Description

Planning

Part 1 : 15/12/2010

  • Development of the GUI
  • Redefine the architecture of the Monteverdi repository to allow the management of new modules and applications based on it

Nothing change? create Sub repositories for Google Earth application's modules?

Part 2 : 15/06/2010

Use cases

Case 1: Automatic process

Data : Charter Product (.jpg?) + (.kml)

  • Process : Automatic Extract Roi (image + legend) -> Affine estimation -> KMZ production
  • TODO:
    • Generic test to validate the procedure (Module without GUI!!)
    • Automatic extract ROI

Case 2:supervised process without orthorectification

  • Data : Charter Product (.jpg)
  • Process : Manual or automatic Extract Roi -> GCP to sensor model (use OSM datas) -> KMZ Production

Case 3:supervised process with orthorectification

  • Data : Charter Product (.jpg)
  • Process : Manual or automatic Extract Roi -> GCP to sensor model (use OSM datas) -> Orthorectification -> KMZ Production

Teh orthorectification process is now realised by the 3D engine of Google Earth!


Dont forget

  • Production of a structured KML products needs a good analysis of the Charter products
  • How introduce temporal informations in the final product?
  • Customized applicaton based on Monteverdi
  • Focus on the need to a simple IO applications for producing "Image to KML"

Monteverdi Modules

Extract ROI

Import Tile Map (OSM)

  • DEV:
    • Freeze if no network interface available -> Search Error: fallback on the origin

NOTE: to prevent the freeze, check if it is possible to use the same method as in the CoordinateToName class: call to the network methods in a separate thread.

  • ENH:
    • PostGIS access
    • Import PNG (need generic approch - not now)
    • Rescale png image (cf mail otb-dev EC). TODO : contribution from C-S -> compression module
    • Replace OSM level by associated cartographic projection (1/20 000...)
    • Asynchrone connection to improve ergonomy

GCP to sensor model

  • ENH:
    • refresh GUI (in progress)
    • Add multiple viewer for OSM


Segmentation

    • Existing tools in gimp to perform palette segmentation?

General Questions/Remarks

  • Google earth proj geographic -> lib proj comme QGIS (lien avec OSSIM discussion à avoir avec EC)
  • streaming données vecteurs VectorDataProperties a la lecture dans le reader -> ??? Is there a size constraint on the vector data? What is the worse case we plan to have?
  • Visu -> hors perimètre car outil n'est pas un outil de visu
  • Cache obliger des données OSM -> forcé le cache
  • ajouter des données vecteur via Postgis -> ???
  • ouverture d'un kml sertit ne fonctionne pas et ne libère pas la données image associée
  • Pertinence d'un outil de recalage vecteur/vecteur -> se servir de la classe ITK ImageDistanceMap dispo cf Examples/Filtering/DanielssonDistanceMapImageFilter.cxx.

Datas

  • Test with real Datas.

With different resolutions, jpg, kml...

    • Sertit :
    • ZKI
    • Haiti SRTM 3

ROADMAP

    • Generic test to validate the procedure
    • KMZ filter to produce a pyramid image (=>OTB contribution, for now it is in the GE-Project)
    • Monteverdi Module for KMZ exportation
    • 2 km errors during exportation to GE (problem precision in lat/long) -> Use Physical Points
    • Production of a KMZ not KML: Need to provide GUI to allow user to integrate images, logos, dates (based on the "Simple Text Editor") example of FLTK
    • Add orthorectification to WGS84 (OTB library)
    • Implement the Case 1 (automatic) KML to KMZ
    • Review of the GCP to sensor model : Estimate a affine transformation + allow to load a list of GCP need to complete
    • Module wizard
    • Automatic segmentation of vector layers
    • Generic process to export datas to Google Earth

Drawbacks

    • Size of associated windows of gcp to sensor model module

Screenshots

  • Extract the region of interest on the VA product (we can imagine an automatic procedure which will extract the ROI and the corresponding legend from the product)

ScreenshotExtractROI.jpg

  • Georeferencing the input VA product with Open Street Map layer (or by import a GCP list from a XML file)

Screenshot GCPToSensorModel.jpg

  • Export the new product as a KMZ file in Google Earth (Note that the legend area appears in a separate file)

Screenshot GE.jpg