robokudo.annotators.image_preprocessor

Image preprocessing and point cloud generation.

This module provides an annotator for:

  • Converting RGB-D camera data to point clouds

  • Handling color and depth image synchronization

  • Managing camera intrinsic parameters

  • Providing visualization modes for debugging

Note

Depth values are expected to be in millimeters.

Classes

ImagePreprocessorAnnotator

RGB-D image preprocessor and point cloud generator.

Module Contents

class robokudo.annotators.image_preprocessor.ImagePreprocessorAnnotator(name: str = 'ImagePreprocessor', descriptor: ImagePreprocessorAnnotator = Descriptor())

Bases: robokudo.annotators.core.BaseAnnotator

RGB-D image preprocessor and point cloud generator.

This annotator:

  • Converts RGB-D camera data to point clouds

  • Handles color/depth image synchronization

  • Manages camera intrinsic parameters

  • Provides visualization modes for debugging

Warning

Requires properly configured camera intrinsics and color-to-depth ratio.

class ViewMode

Visualization mode enumeration.

color: int = 1

Display color image (1)

depth: int = 2

Display depth image (2)

class Descriptor

Bases: robokudo.annotators.core.BaseAnnotator.Descriptor

Configuration descriptor for image preprocessing.

class Parameters

Parameters for configuring preprocessing.

depth_trunc: float = 9.0

Maximum depth distance in meters, points beyond are discarded from cloud creation

parameters
color: typing_extensions.Optional[numpy.typing.NDArray] = None

Optional color image currently being worked with

depth: typing_extensions.Optional[numpy.typing.NDArray] = None

Optional depth image currently being worked with

display_mode: int = 1

The display mode to use (color or depth)

update() → py_trees.common.Status

Process RGB-D images and generate point cloud.

The method:

  • Loads color and depth images from CAS

  • Synchronizes and scales images

  • Converts to Open3D format

  • Generates point cloud

  • Updates visualization based on view mode

Returns:

SUCCESS after processing

Raises:

RuntimeError – If color-to-depth ratio is not set

key_callback(key: int) → None

Handle keyboard input for view mode switching.

Parameters:

key – ASCII value of pressed key

Available keys:

  • ‘1’: Switch to color view mode

  • ‘2’: Switch to depth view mode