The robot arm controller is a Raspberry Pi 2 Model B. The Servomotors are Dynamixel AX-12A. There is a Raspberry Pi camera module mounted on the top for image processing.

The Computer Vision algorithms applied here are Edge Detection, Binarization, Pixel Expansion, Labeling and Object Extraction. In this Video I tried to show how the robot see’s the world by adding pictures directly out of the Image Processing algorithms (I just added the coloring in the Labeling process). I also tried to sync the pictures to the superb music of the great artist “broke for free”.

Here's some further info on the thing:

I didn’t use OpenCV. The image processing algorithms applied here are all very simple. I wanted to write them by my own. Two important libraries which I used are pythons "picamera" and a library called "ax12". "picamera" provides an easy way to get greyscale pixeldata from the Raspberry Pi camera module. "ax12" is used for the communication with the Dynamixel AX-12A servos. I did write some code to make the servomotors move smoother (starting and stopping in a smooth sinusoidal manner). And then there was a bit of code to actually get the junctions into positions, which would allow the electromagnet to pick up the metallic things. In other words, this was about getting the thing to move correctly given some x and y values which were extracted from the image earlier.

Music:
"Night Owl" by Broke For Free

свернуть