Automatic Path-Planning Algorithm for Realistic Decorative Robotic Painting

The industrial spray-painting process is currently split into two main branches: the uniform covering of surfaces (e.g. automotive), and decorative painting (e.g. furniture, porcelain). While the former makes extensive use of robotics, the latter does not, relying instead on the manual skills of expert painters. In this paper we propose an innovative algorithm to reproduce non-uniform, photorealistic, grey-scale images on surfaces, using an ordinary industrial spray- painting robot. The algorithm splits the process into a set of iterative steps with the spray gun located at decreasing distance from the canvas or generic surface. Since the spray morphology is cone-shaped, at higher distances the stroke will be bigger than at closer distances. Taking advantage of this characteristic, the algorithm can efficiently build up the image starting with a big stroke to paint large details of the image. Then, with increasingly smaller strokes, it can paint the rest of the smaller details. The core is based on a constrained ordinary least squares (OLS) method. The image is segmented and a path is defined using an approach normally used in CAM: offset generation via Voronoi diagram. A set of critical points in the image are then chosen to avoid oversaturation and used to build a system which is then solved using OLS. Depending on the image to reproduce, this methodology promises to be far more efficient than printer or plotter-like painting processes, where the image is built one point at a time. Applications of this innovative methodology are range in the field of decorative or functional painting, from furniture to houses and buildings, to automotive, to naval architecture.