Skip to content

Global Predefined Parameters 'global_with_depth' and 'global_with_visualisation' for the Annotators

Max Philipp Gandyra requested to merge dev_anno_params into main

Using the 'self.descriptor.parameters.global_with_depth' value in some annotator instance returns:

  • the global predefined value of 'robokudo.annotator_parameters.AnnotatorPredefinedParameters.global_with_depth', if 'global_with_depth' is/was not defined in the annotator 'Parameters' instance/class (e.g. 'self.global_with_depth = False')
  • or the local value otherwise.

The global 'global_with_depth' value can be directly accessed and/or changed with e.g. 'robokudo.annotator_parameters.AnnotatorPredefinedParameters.global_with_depth = True'. This value controls whether the depth image and derived data from it, e.g. point cloud, should be used by the annotators. For example 'CASViews.DEPTH_IMAGE = None' should no longer causes errors in yolo or similar annotators if 'global_with_depth = False'. This value should normally be set by either in the pipeline definition or in the camera interface.

The global 'AnnotatorPredefinedParameters.global_with_visualisation' value has the same read/write properties as 'global_with_depth' and controls whether the annotators should create a visual representation for their results. Setting it to "False' should save computation resources and time, if no visualization is needed. In comparison, RoboKudo 'headless mode' still creates the visual representations, but does not display them.

Edited by Max Philipp Gandyra

Merge request reports