Customer: Centre National d’Études Spatiales (CNES)

Programme: Generic

Supply Chain: CNES > CS Group SPACE

Context

CS Group responsabilities for Orfeo Toolbox are as follows:

• Maintenance and library developments
• Community assistance
• Animation of user and developer communities
• Maintenance of the platform for online services

The features are as follows:

• Segmentation **: Applicable on a large scale (Mean Shift, Watersheds, etc.) Vectorization, comparison **Machine Learning **: Training of classification and regression models. Application on images and vectors data. Assessment **Primitive extraction **: contours detection, calculation of radiometric indices, texture extraction, calculation of global and local statistics, mathematical morphology **Image geometry **: ortho-rectication, model refining, re-sampling on grid, pansharpening, disparity calculation, image rechocation **Calibration, hyperspectral …

Project implementation

The project objectives are as follows:

• **OTB **: Free image processing processing toolbox for remote sensing, composed of a C ++ library and a set of applications accessible from different Wrappers: Python 3, command line, QGIS plugin, Gui.

The processes for carrying out the project are:

• Iterative development
• Continuous development and integration (coding/unit testing)
• Automatic Test Plan Generator with test execution and insertion of results in a CDash instance

Technical characteristics

The solution key points are as follows:

• C ++ software based on OTB libraries
• Python
• CMAKE / CTEST / CDASH
• Management of several plugins using generic factories

The main technologies used in this project are:

Domain	Technology(ies)
Operating System(s)	Linux, Windows
Programming language(s)	C++ 14, Python
Production software (IDE, DEVOPS etc.)	CMake, SWIG, Gitlab, Gitlab-CI, Sphinx, Discourse, Wordpress, Doxygen
Main COTS library(ies)	ITK, GDAL, Boost, 6S, MuParser, Shark ML