Summer internship: Real-time object tracking and colour sorting

My internship in the Smart Systems research group has now come to an end, and the final month turned out to be perhaps the most productive.

I printed and mounted the new gripper jaws that I mentioned in my previous post. The robot can now grasp wider objects, and small positioning deviations no longer affect the pick-up as much. I also taught the robot to handle tilted objects: using the RANSAC algorithm, it estimates the surface angle and adapts the gripper orientation accordingly.

I also implemented real-time object tracking and colour sorting using my Python library and documented how to use these features in the project guide.

As a final demonstration, I made a short video in which the robot detects objects of a selected colour one by one, picks them up, and decides where to place them in a row.

I also had the chance to demonstrate the system live during a presentation for new students in the Mechanical Engineering department. It was great to present the work, answer questions, and show what can be done with the robot. Maybe some of these students will one day use the guide when working with the system themselves.

Over these three months, I have learned a lot, from soldering wires to implementing computer vision algorithms. I am very grateful for the opportunity to work on this project and for the support of my supervisor, Wallace Moreira Bessa, and my fellow researchers, Gabriel da Silva Lima and Kimmo Paldanius.

And who knows, perhaps this work will continue into a deeper research project and eventually become the topic of my master’s thesis.

You can see a short demonstration of the real-time object tracking, colour sorting and adaptive grasping here: [YouTube]