Talk:Supervised classification (Tutorial)

From AWF-Wiki
Jump to: navigation, search


Image statistics(not mandatory, only if normalization of the variables is required)

  • Type into the search box of the Windows taskbar: mapla.bat. Click on mapla.bat to open Monteverdi Application Launcher.
  • In the search engine of mapla, type ComputeImagesStatistics and double click ComputeImagesStatistics.
  • Specify a multispectral image as Input Image: the Sentinel-2 image Subset_S2A_MSIL2A_20170619T_MUL_BOA.tif
  • Specify directory and name for the XML Output image. Specify the extension .xml for this file.
  • Click on Execute.

Qgis ComputeImagesStatistics.png

Classification with Orfeo Toolbox (Outdated Qgis 2)

Image statistics

  1. Add the Sentinel-2 imagery Subset_S2A_MSIL2A_20170619T_MUL_BOA.tif into a QGIS project.
  2. Calculate mean and standard error for each band of the Sentinel-2 imagery using the OTB Graphical User Interface.
  • In the Search box on the Windows Start menu type OSGeo4W Shell. You should be able to open the shell by clicking on it.
  • Type into the shell: otbgui_ComputeImagesStatistics.

Select a multiband input file and an output XML file as seen in the screenshot below. Qgis ComputeImagesStatistics.png

Train image classifier

  1. Add the training areas as vector polygon file lab05_training_input.shp into QGIS.
  2. Open Orfeo Toolbox --> TrainImageClassifier (libsvm) to use the Support Vector Machine SVM algorithm (see figure B).
  3. Set Subset_S2A_MSIL2A_20170619T_MUL_BOA.tif as Input image list.
  4. Set lab05_training_input.shp as Input vector list.
  5. Set Subset_S2A_MSIL2A_20170619T_MUL_BOA.tif.xml as Input XML image statistics file.
  6. Set Name of discrimination field to C_ID (C_ID refers to the column that contains the LUC code).
  7. Save the Output confusion matrix as ConfusionMatrixSVM.csv.
  8. Save the Output model as SVM.model.
  9. Calculation of accuracies :
    Open ConfusionMatrixSVM.csv in LibreOffice or MS Excel and calculate overall, producer and consumer accuracies.


  1. Open Orfeo Toolbox --> Image Classification (see figure C).
  2. Set Subset_S2A_MSIL2A_20170619T_MUL_BOA.tif as Input image.
  3. Set SVM.model as Model file.
  4. Set Subset_S2A_MSIL2A_20170619T_MUL_BOA.xml as Statistical file.
  5. Save the Output image as su_svm.tif.
  6. Evaluate classification results.
    1. Add the classification result su_svm.tif to QGIS.
    2. Right click su_svm.tif in the TOC and select Properties --> Style --> Style --> Load Style.
    3. Load lab05_MinDist.qml.

Compute a confusion matrix with independent reference data

  1. Open Orfeo Toolbox --> ComputeConfusionMatrix (Vector).
  2. Set su_svm.tif as Input image.
  3. Set lab05_validation.shp as Input reference vector data.
  4. Set Field name to C_ID.

GRASS implementation

When testing for GRASS in the Semi-Automatic Classification Plugin's Settings tab, I needed to have grass set up first (i.e. created a grassdata folder with a location and mapset); this makes sense, but I don't know if we provide a tutorial for this yet? - Levent (talk) 17:22, 24 February 2014 (CET)

This Module does not work in QGIS 2.18-ltr:

    1. Open Orfeo Toolbox --> Compute images second order statistics (see figure A).
    2. Set Subset_S2A_MSIL2A_20170619T_B12_BOA.tif as Input images.
    3. Save the Output XML file as Subset_S2A_MSIL2A_20170619T_B12_BOA.tif.xml.

Outdated figures:


Figure A: Dialogues of the Compute images second order statistics
Figure B: Dialogues of the TrainImageClassifier (libsvm)
Figure C: Dialogues of the Image Classification
Figure D: Dialogues of the ComputeConfusionMatrix (Vector)
Personal tools