Visual Servoing to Sort Objects With Baxter Using ROS
Overview:
- Making the Baxter robot pick green and red blocks and place them in their respective boxes.
- The Services framework of ROS was used for communicating between nodes, MoveIt! and TRAC-IK were used for motion planning, and OpenCV image processing was done on the left hand camera feed to detect objects.
- Most of my contribution was using OpenCV to process the left-hand camera feed (thresholding, morphological operations, finding centroids of contours), and experimentally finding the camera calibration factor to convert pixel coordinates to Baxter’s base frame coordinates.
- This was the final project of the course Embedded Systems in Robotics, taught by Dr. Jarvis Schultz at Northwestern University. Team members: Jose Miranda, Sofya Akhmametyeva, Yves Nazon and Tanay Choudhary
- For the complete documentation and code, head to the GitHub page here
Baxter Sorting Objects by Color from Tanay Choudhary on Vimeo.
Project Dependencies:
Project Details
Date: Dec 13, 2015
Categories: project
