VectorData and Image superimposition status

This page tries to summarize the different scenario which can appear with the different type of image and vector data managed by OTB/Monteverdi.

The different type of image managed by OTB/Monteverdi can be classify in 4 groups:

• Image with geographic coordinates (for example image with projection system = WGS84) => IMG_geo
• Image with cartographic coordinates (for example image with projection system = WGS84/UTM) => IMG_carto
• Image with sensor model (need to have a keyword list) => IMG_sm
• Image with index coordinates (no projection Ref and no keyword list) => IMG_ind

The different type of vector data managed by OTB/Monteverdi can be classify in 3 groups:

• Vector data with geographic coordinates (for example vector data with projection system = WGS84) => VD_geo
• Vector data with cartographic coordinates (for example vector data with projection system = WGS84/UTM) => VD_carto
• Vector data with index coordinates => VD_ind

We can notice that OTB/Monteverdi managed an other type of vector data but this one cannot be saved with all its information in a classical shape file format. Indeed this vector data object can support, inside OTB/Monteverdi, a keyword list, an origin and a spacing. Each previous type of vector data can embedded this information if it was computed by a filter from OTB library.

Visualization cases

Typical use: open and view image + vector data

Open an (Geo/Carto/Sensor Model/Index) image and vector data (Geo/Carto/Index) and try to view these dataset with the viewer.

We should be able to visualize all the different cases except cases with IMG_ind and VD_carto/VD_geo. This two cases should produce at least a warning message.

Superimposition of VD_ind and all images are possible only in the coordinate system of the layer image.

Using OTB vector data

When we try to visualize VD_geo or VD_carto with (Geo/Carto/Sensor Model) image viewer module behavior should be the same as the typical case. We didn't take into account the additional information. It is the same for the IMG_ind case, although the new meta-data in the vector data, it is not possible to superimpose physical coordinate on it.

Add keyword list, origin and spacing to a VD_index give the possibility to consider, inside OTB pipeline, this dataset as a vector data with a sensor model. With this type of dataset, it is possible to manipulate physical coordinate.

Vectorization cases (scenario)

Remark: type of vector data generated by the vectorization module is based on the type of input image.

Scenario 1:

• Open an (Geo/Carto/Sensor Model/Index) image.
• Launch the vectorization module with this image.
• Generate a vector data.
• Close the vectorization module.
• Connect the vectorization module output to the viewer module with the same image or with another image type.

Scenario 2:

• Open an (Geo/Carto/Sensor Model/Index) image.
• Launch the vectorization module with this image.
• Generate a vector data.
• Close the vectorization module.
• Reopen the vectorization module with the same image or an different input image.
• Close the vectorization module.
• Add a polygon.
• Connect the vectorization module output to the viewer module with the original image, the second image or with another image type.

Scenario 3:

• Open an (Geo/Carto/Sensor Model/Index) image.
• Make an extract of this image
• Launch the vectorization module with this image.
• Generate a vector data.
• Close the vectorization module.
• Connect the vectorization module output to the viewer module with the original image, the extract or with another image type.

Scenario 4:

• Open an (Geo/Carto/Sensor Model/Index) image.
• Launch the vectorization module with this image.
• Generate a vector data.
• Close the vectorization module.
• Save the vector data as a shape file.
• Reopen the vector data.
• Connect the vector data reader module output to the viewer module with the same image or with another image type.

Scenario 5:

• Open an (Geo/Carto/Sensor Model/Index) image.
• Make an extract of this image
• Launch the vectorization module with this image.
• Generate a vector data.
• Close the vectorization module.
• Save the vector data as a shape file.
• Reopen the vector data.
• Connect the vector data reader module output to the viewer module with the original image or the extract or with another image type.

Complex scenario

This scenario represents a complete pipeline which exposes the possibility of OTB about semi-supervised classification.

• Open an (Geo/Carto/Sensor Model/Index) image.
• Make an extract of this image
• Launch the vectorization module with this image / or a filter which generate vector data from an image.
• Connect the classification module (Experimental) or the list sample generator + SVM classifier filter.
• Check the labeled output.