OTB Modules

Since OTB 9.0, the toolbox is split in packages that can be installed independently. This table shows the list of modules :

Module Name

Criticity

Core and Dependencies

Mandatory

FeaturesExtraction

Optional

Learning

Optional

StereoProcessing

Optional

SAR

Optional

Hyperspectral

Optional

Core and Dependencies

These are the two only mandatory modules for OTB.

OTB-Core contains optical image preprocessing applications (Pansharpening, Calibration, Orthorectification, etc.), most useful image manipulation tools (ExtractROI, Superimpose, Mosaic, Concatenate, ..) and various applications to manipulate vector data. OTB-Dependencies contains all third-parties libraries needed to execute Core / All applications

For the dependencies, there are two packages available depending on your needs :

  • OTB-All-Dependencies this is the package that you should install by default : contains the dependencies for the core and optional modules.If you install the Core and decide to use an optional module afterwards, you just need to install this optional module over the current OTB install.

  • OTB-Core-Dependencies suits users who only need the core fonctionnalities of OTB.If you plan to install an optional module afterwards, you need to download and install the “All-Dependencies” package before installing the module.

The application list for each package is described in the ::doc::Applications` section

FeaturesExtraction

This optional module contains classical filtering applications, such as texture extraction, edge extraction, smoothing, morphological operations. This module relies on external libraries (MuParser / MuParserX) contained in OTB-All-Dependencies. it can be installed along with the Core package with a simple tar extract command and directly available after sourcing the otbenv.profile

Learning

This module contains Machine Learning applications based on classical supervised or unsupervised algorithms (SVM, Random Forest, K-Means, Multi-layer Perceptron Neural Network, etc.). Its applications handle the whole processing chain : sample selection, learning, prediction and finalization of a classification map. It also contains a regression framework.

StereoProcessing

This module allows to register stereoscopic images and compute disparity in order to build a 3D model from a couple of satellite images in raw / sensor geometry.

SAR

This module contains SAR (Synthetic Aperture Radar) applications and covers all the scope of SAR imagery manipulation : SAR Calibration, Unspeckle with temporal filtering, Interferogram, etc.

Hyperspectral

This module contains specific applications to handle hyperspectral images : spectral angle classification, end member number estimation or unmixing.