A year after designing our simple line follower robot, we decided to go further and add more complex features! We started with replacing IR sensors with a camera. We also changed the design altogether. We wanted our robot to differentiate between lines and shapes of different colors. So we started to learn image processing! Here is what we build:
The overall procedure is described as follows :
- First, we sent the camera stream through a wired connection to our pc and then processed the image with OpenCV in C++. We also tested an IP camera and captured frames through a local network that worked in real-time.
- After processing the image, we sent proper controlling commands to the robot with a USB to TTL module. We also experimented with HMT-HMR modules.
This project was our first experience with computer vision and we did not use any ML/DL techniques. We created robust and custom color filters, and we were able to detect shapes through hand-crafted features like the number of corners and angles between lines!