The image_pipeline stack is designed to process raw camera images into useful inputs to vision algorithms: rectified mono/color images, stereo disparity images, and stereo point clouds. Components include:

  • Calibration: Cameras must be calibrated in order to relate the images they produce to the three-dimensional world. The camera_calibration package provides tools to calibrate monocular and stereo cameras in your ROS system. A CameraInfo wiki page provides a detailed description of the parameters used by the pipeline.

  • Monocular processing: The raw image stream can be piped through the image_proc node to remove camera distortion. The node also performs color interpolation for Bayer pattern color cameras.

  • Stereo processing: The stereo_image_proc node performs the duties of image_proc for a pair of cameras co-calibrated for stereo vision. It also uses stereo processing to produce disparity images and point clouds.

  • Depth processing: depth_image_proc provides nodelets for processing depth images (as produced by the Kinect, time-of-flight cameras, etc.), such as producing point clouds.

  • Visualization: The image_view package provides a lightweight alternative to rviz for viewing an image topic. It also includes a stereo_view tool for viewing stereo pairs and disparity images.

Hardware Requirements

The image pipeline will work with any conforming ROS camera driver node. See camera_drivers for cameras already supported in ROS. The minimal requirements for such a node are:

Generic camera node

Published Topics

image_raw (sensor_msgs/Image)
  • The unprocessed image data.
camera_info (sensor_msgs/CameraInfo)
  • Contains the camera calibration (if calibrated) and extra data about the camera configuration.


set_camera_info (sensor_msgs/SetCameraInfo)

Camera drivers normally have parameters (exposure, gain, etc.) that you can configure at runtime. stereo_image_proc has runtime-configurable stereo processing parameters of its own. The rqt_reconfigure is useful for tweaking your configuration to get best results.

Report a Bug

<<TracLink(ros-pkg image_pipeline)>>

Wiki: image_pipeline (last edited 2013-10-30 08:33:30 by IsaacSaito)