A software pipeline that converts an input image into drawable robot trajectories, connecting image processing, path planning, and G-code generation to a physical robotic artist.
My contribution to the Robot Artist project was the software pipeline that takes an input image, detects the geometry worth drawing, organizes that geometry into paths, and converts those paths into G-code the robot can follow.
Edge detection produced the geometry of the image, but the robot still needed an ordered trajectory to follow. The main challenge was bridging that gap: taking the detected edge information and organizing it into paths that could be translated into continuous robot motion and, ultimately, G-code.
The project demo below brings the process together, including the software workflow, simulation, and the physical robot producing a drawing.
This project gave me experience building the link between an abstract image-processing result and real robot motion. The key challenge was not simply finding edges, but turning those edges into an ordered representation that could ultimately be expressed as executable machine motion.