Courses / programming

ROS 2 Perception

advanced~18 hours13 lessonsEnvironment: docker

Prerequisites: ros2-advanced, gazebo-sim

Start: Image topics and cv_bridgePractice questions

Modules

Module 1 — Cameras in ROS
  1. 1Image topics and cv_bridge
  2. 2Camera calibration and intrinsics
  3. 3Compressed transport
Module 2 — Classical vision
  1. 1OpenCV in a ROS node
  2. 2Color filtering and contours
  3. 3ArUco marker detection and pose
Module 3 — Point clouds
  1. 1Depth cameras and lidar clouds
  2. 2PointCloud2 and processing pipelines
  3. 3Filtering, downsampling, clustering
Module 4 — Learning-based detection
  1. 1Running an object detector in ROS
  2. 2From bounding box to 3D position
  3. 3Publishing detections into TF
  4. 4Project: detect-and-locate objects on a table

Textbooks

Video resources

  • Camera and photogrammetry lectures — Cyrill Stachniss