project logo

OTB Quickstart

ORFEO Toolbox library (OTB) is a high performance library for image processing targeted on remote sensing.

This Quick Start describes how to:

  • Get metadata informations in an image
  • Perform mathematical operations between image bands
  • Open raster images with the application monteverdi, perform segmentation (mean-shift clustering) and visualize the result
  • Perform supervised classification based on Support Vector Machine algorithm

The OTB applications provide lot’s of interesting tools which facilitate the manipulation of images. All these tools are available through:

  • CLI : command line interface
  • GUI : in a standalone graphical user interface (in Qt)
  • QGIS plugin : available through Sextante

See

Sample data used in this quickstart can be found here:

Display metadata informations in an image

You can get all the metadata informations contained in an image with the command : otbcli_ReadImageInfo The unique parameter is the Input image file name, for example : otbcli_ReadImageInfo -in qb_RoadExtract.tif

Calculator on image bands

The otbcli_otbBandMath provides an efficient way to perform mathematical operation on monoband images. The syntax is quite simple, for example substrating two bands to study the image differences on the images SpotBefore.tif and SpotAfter.tif, just use the command : otbcli_BandMath -il SpotBefore.tif SpotAfter.tif -out difference.tif -exp “im1b1-im2b1” The application is able to perform complex mathematical operations over images (threshold, logarithmic rescaling...). This homebrewed digital calculator is also bundled with custom functions allowing to compute a full expression. For example, as remote sensing images measure physical values, it is possible to extract several indices with physical meaning like the NDVI (Normalized Difference Vegetation Index) for the vegetation. With the calculator you’re able to compute the NDVI on a multispectral sensors images by doing: otbcli_BandMath -il qb_RoadExtract.tif -out ndvi.tif -exp “ndvi(im1b3,im1b4)”

Pixel based classification

The classification in the application framework provides a supervised pixel-wise classification chain based on learning from multiple images, and using one specified machine learning method like SVM, Bayes, KNN, Random Forests, Artificial Neural Network, and others...(see application help of TrainImagesClassifier for further details about all the available classifiers). It supports huge images through streaming and multi-threading. The classification chain performs a training step based on the intensities of each pixel as features. Please note that all the input images must have the same number of bands to be comparable.

Perform segmentation with Monteverdi

  • Start Monteverdi from its icon from the Spatial Tools folder on the desktop

  • Select an raster image, using File ‣ Open Dataset ‣ /home/user/otb/qb_RoadExtract.tif

  • Go to the Filtering ‣ Mean Shift clustering

  • Select the input raster image (Reader0) from the input window selection

  • Verify you can tune parameters of the segmentation and see the result on the region of interest by clicking on “Run”

  • Select “Close” when you are satisfied by the result.

  • In the main window, right click on the “Clustered Image” in the resulting dataset “MeanShift0” and select “Display in viewer”

    ../../_images/otb-mean_shift.jpg

Perform supervised classification based on SVM with Monteverdi

  • Start Monteverdi from its icon from the Spatial Tools folder on the desktop

  • Select an raster image, using File ‣ Open Dataset ‣ /home/user/otb/qb_RoadExtract.tif

  • Go to the Learning ‣ SVM classification

  • Select the input raster image (Reader0) from the input window selection

  • You can add classes (Add Class button), select learning samples by drawing polygons in the

  • Go to the Setup ‣ SVM to set the classification algorithm parameters

  • Click on the Learn button to create a classification model fron the input learning classes

  • Click on the Display button to show the result of the supervised classification on the entire image

    ../../_images/otb-svm.jpg

For the full tutorial see the article.

What Next?

  • OTB Software Guide

    The main source of information is the OTB Software Guide. This is a comprehensive guide which comprises about 600 pages, detailing the steps to install OTB and use it. Most of the classes available are heavily illustrated with results from real remote sensing processing.

  • OTB CookBook

    A guide for OTB-Applications and Monteverdi dedicated for non-developers is also available.This guide is composed of a brief tour of of OTB-Applications and Monteverdi, followed by a set of recipes to perform usual remote sensing tasks with both tools.

  • OTB Tutorials

    Follow the tutorials to learn more about OTB.

  • OTB Applications documentation

    See also detailed documentation about OTB applications

  • OTB courses with Pleiades images

    Follow the courses to learn more about OTB.

Copyright & Disclaimer