Commit 57493726 authored by Guillaume Pasero's avatar Guillaume Pasero

DOC: python recipe on new services

parent 2ef5e7f0
......@@ -181,22 +181,132 @@ Interactions with OTB pipeline
The application framework has been extended in order to provide ways to
interact with the pipelines inside each application. It applies only to
applications that use input or output images. Let's check what are the
functions added to the ``Application`` class:
* ``GetImageOrigin(key, idx=0)``: returns the origin of the image (physical position of the first pixel center)
* ``GetImageSpacing(key, idx=0)``: returns the signed spacing of the image
* ``GetImageSize(key, idx=0)``: returns the size of the LargestPossibleRegion
* ``GetImageNbBands(key, idx=0)``: returns the number of components per pixel
* ``GetImageProjection(key, idx=0)``: returns the ProjRef
* ``GetImageKeywordlist(key, idx=0)``: returns the Ossim keywordlist
* ``GetImageMetaData(key, idx=0)``: returns the itk::MetaDataDictionary
* ``GetImageRequestedRegion(key, idx=0)``
* ``PropagateRequestedRegion(key, region, idx=0)``: set a given RequestedRegion on the image and propagate it, returns the memory print estimation
* ``GetImageBasePixelType(key, idx=0)``: helper function to guess the pixel type of the underlying Image/VectorImage.
* ``GetParameterImageBase(key, idx=0)``: helper, returns the image as a ImageBaseType (works on all image parameters)
* ``SetParameterImageBase(key, idx=0)``: helper, sets the parameter InputImage/InputImageList/ComplexInputImage with an ImageBaseType
Note: some of these function use an optional index to refer to an image parameter. By default this index is 0. It allows to access InputImageLists transparently.
functions added to the ``Application`` class. There are a lot of getter
| Function name | return value |
| ``GetImageOrigin(...)`` | origin of the image (physical position|
| | of the first pixel center) |
| ``GetImageSpacing(...)`` | signed spacing of the image |
| ``GetImageSize(...)`` | size of the LargestPossibleRegion |
| ``GetImageNbBands(...)`` | number of components per pixel |
| ``GetImageProjection(...)`` | Projection WKT string |
| ``GetImageKeywordlist(...)`` | Ossim keywordlist (sensor model) |
| ``GetImageMetaData(...)`` | the entire MetaDataDictionary |
| ``GetImageRequestedRegion(...)``| requested region |
| ``GetImageBasePixelType(...)`` | pixel type of the underlying |
| | Image/VectorImage. |
All these getters functions use the following arguments:
* ``key``: a string containing the key of the image parameter
* ``idx``: an optional index (default is 0) that can be used to access ImageList
parameters transparently
There is also a function to send orders to the pipeline:
``PropagateRequestedRegion(key, region, idx=0)``: set a given RequestedRegion
on the image and propagate it, returns the memory print estimation. This function
can be used to measure the requested portion of input images necessary to produce
an extract of the full output.
Note: a requested region (like other regions in the C++ API of otb::Image) is
just a pair of an image index and a size, that define a rectangular extract of
the full image.
This set of function has been used to enhance the bridge between OTB images
and Numpy arrays. There are now import and export functions available in
Python that preserve the metadata of the image during conversions to Numpy
* ``ExportImage(self, key)``: export an output image parameter into a Python
* ``ImportImage(self, key, dict, index=0)``: import the image from a Python
dictionary into an image parameter (as a monoband image).
* ``ImportVectorImage(self, key, dict, index=0)``: import the image from a
Python dictionary into an image parameter (as a multiband image).
The Python dictionary used has the following entries:
* ``'array'``: the Numpy array containing the pixel buffer
* ``'origin'``: origin of the image
* ``'spacing'``: signed spacing of the image
* ``'size'``: full size of the image
* ``'region'``: region of the image present in the buffer
* ``'metadata'``: metadata dictionary (contains projection, sensor model,...)
Now some basic Q&A about this interface:
Q: What portion of the image is exported to Numpy array?
A: By default, the whole image is exported. If you had a non-empty requested
region (the result of calling PropagateRequestedRegion()), then this region
is exported.
Q: What is the difference between ImportImage and ImportVectorImage?
A: The first one is here for Applications that expect a monoband otb::Image.
In most cases, you will use the second one: ImportVectorImage.
Q: What kind of object are there in this dictionary export?
A: The array is a numpy.ndarray. The other fields are wrapped
objects from the OTB library but you can interact with them in a
Python way: they support ``len()`` and ``str()`` operator, as well as
bracket operator ``[]``. Some of them also have a ``keys()`` function just like
This interface allows you to export OTB images (or extracts) to Numpy array,
process them by other means, and re-import them with preserved metadatas.
Here is a small example of what can be done:
.. code-block:: python
import otbApplication as otb
# Create a smoothing application
app = otb.Registry.CreateApplication("Smoothing")
# only call Execute() to setup the pipeline, not ExecuteAndWriteOutput() which would
# run it and write the output image
# Setup a special requested region
myRegion = otb.itkRegion()
myRegion['size'][0] = 20
myRegion['size'][1] = 25
ram = app.PropagateRequestedRegion("out",myRegion)
# Check the requested region on the input image
# Create a ReadImageInfo application
app2 = otb.Registry.CreateApplication("ReadImageInfo")
# export "out" from Smoothing and import it as "in" in ReadImageInfo
ex = app.ExportImage("out")
app2.ImportVectorImage("in", ex)
# Check the result of ReadImageInfo
someKeys = ['sizex', 'sizey', 'spacingx', 'spacingy', 'sensor', 'projectionref']
for key in someKeys:
print(key + ' : ' + str(app2.GetParameterValue(key)) )
# Only a portion of "out" was exported but ReadImageInfo is still able to detect the
# correct full size of the image
Corner cases
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment