The authors present a multi-color robotic pen drawing system, TSP-Bot, that converts digital raster images into continuous paths that can be drawn on a physical canvas using robotic manipulators. The key steps are:
Color Processing: The input image is split into predefined color channels, such as CMYK, to enable reproduction using a limited color palette.
Stippling: A stippling algorithm is used to generate a set of points that represent the tone of the original image for each color channel.
TSP Solving: A Traveling Salesman Problem (TSP) solver is used to find a piecewise-continuous path that visits each stippled point exactly once for each color channel.
Path Optimization: The piecewise-linear TSP path is simplified and smoothed using the Ramer-Douglas-Peucker algorithm and cubic Bézier spline curves with bounded curvature to ensure smooth robot motion.
Robotic Curve Rendering: The optimized drawing path is mapped to the robot's configuration space using inverse kinematics to execute the drawing task.
The authors demonstrate the system's capabilities using two different robotic setups: a dual-arm manipulator with an automated pen-changing mechanism and a mobile manipulator. The results show intricate, color-accurate TSP pen art drawings created by the robotic system.
A otro idioma
del contenido fuente
arxiv.org
Ideas clave extraídas de
by Daeun Song,E... a las arxiv.org 04-12-2024
https://arxiv.org/pdf/2210.07592.pdfConsultas más profundas