Pixel synchronous measurement of object shape and colour

In this paper we present an extension to high resolution multi-view fringe projection systems, as introduced in [1], in order to add pixel synchronous colour information to 3D point clouds. Our approach allows the use of a separate colour camera without the drawback of having to rely on complex image-tosurface mapping algorithms. The advantage of the new method is that no loss of spatial measurement information occurs. Additionally, colour measurement is not limited like in previous coloured structured lighting approaches. Furthermore this paper covers the task of seamless fusion of coloured multi-view 3D point clouds. Due to differences in lighting and viewing direction in each single object view a simple data fusion causes visual artefacts and seams in the final object illustration. Several approaches to this problem have been applied and evaluated. This includes point-bypoint lighting and viewing angle normalization via physics-based reflection models and measurement geometry aided image processing algorithms. The result of these considerations is an automatic system to create the point cloud of a 360-degree object view with each measuring point containing range and colour information