The site is read-only. Please transition to use Robotics Stack Exchange
| ROS Resources: Documentation | Support | Discussion Forum | Index | Service Status | ros @ Robotics Stack Exchange |
 | 1 | initial version |
The Nao's cameras work just like any other monocular camera. But what do you mean with "calculate real world coordinate"? Without any additional information (like a ground plane) you cannot derive 3D coordinates from 2D image coordinates. For that you need stereo vision which the Nao's camera's can't provide. Monocular SLAM (visual odometry, structure from motion, ...) could be a solution for that, but is a research topic of its own.
You can still calibrate your Nao's cameras just like any other webcam, it only lacks a proper ROS-enabled driver that outputs the correct topics for the automatic calibration. I would suggest to record 10 static frames with the checkerboard in various positions and then feed them into the OpenCV camera calibration.
ROS Answers is licensed under Creative Commons Attribution 3.0 Content on this site is licensed under a Creative Commons Attribution Share Alike 3.0 license.
