.. _edge_detection:

Edge Detection
==============

This pipeline produces an edge map from a single image using the MoGe2 model.

.. warning::

  This pipeline is only available if the `CARS Edge Detection plugin <https://github.com/CNES/cars-edge-detection-plugin>`_ is installed.

Allowed inputs
--------------

Subsampling can only take sensor images as inputs, as shown in the :ref:`input <input>` section of the documentation.
While the pipeline computes the edge map from a single image, it is necessary to provide images as pairs, as is the case for the other pipelines in CARS.

Applications
------------

This pipeline uses the following applications : 

- :ref:`depth_map_generation <depth_map_generation_app>`

Advanced Parameters
-------------------

.. list-table::
    :header-rows: 1

    * - Name
      - Description
      - Type
      - Default value
    * - save_intermediate_data
      - Save intermediate data for all applications inside this pipeline.
      - bool
      - False
    * - right_image_edge_detection
      - Whether to run edge detection on the right images as well.
      - bool
      - False

When save_intermediate_data is set to True, the following intermediate data will be saved for all processed images:

- ``depth_map.tif``: the depth map generated by the MoGe2 model.
- ``normals.tif``: the normal map generated by the MoGe2 model.
- ``tile_id.tif``: the map specifying the tile ID of each section, generated as a result of tiled processing.