Currently, photonic mixing device (PMD) cameras undergo a great deal of attention. They allow simultaneous recordings of amplitude and distance images with one shot. This opens up new application possibilities like drivers’ assistance in vehicles or gesture control in the multimedia sector. Unfortunately, PMD cameras reach only low spatial resolution. Wherein the pixel resolution for state-of-the-art indoor cameras ranging about VGA resolution, they are even lower for outdoor applications. This limits the possibilities for object recognition. From two-dimensional (2D) imaging there are already methods known for increasing spatial resolution virtually. It means resolution enhancement without changing physically given sensor specifications like pixel dimension or sensor size. In this context, often referred as superresolution (SR). This work compares four well-known geometric SR algorithms from 2D imaging adapted to PMD imaging. Resolution enhancement and quality of the SR results are evaluated objectively by measuring the spatial frequency response (SFR) and investigating the noise performance in amplitude and distance images. Based on these results, SR algorithms for possible measurement tasks in metrological or photographic applications are proposed.
Three-dimensional image acquisition is still a growing field in optical metrology. Various methods are available to reconstruct an object’s three-dimensional surface. The five main types of 3D cameras are stereo cameras, triangulation (pattern or laser scanning), interferometry, light-field cameras and ToF (time-of-flight) cameras. PMD (photonic mixing device) cameras measure the time of light, and thus belong to the field of ToF cameras. Each camera type has fields of application for which it is particularly well suited. Even within PMD cameras, there is a distinction made between applications for indoor and outdoor use. Until today, there is no method to measure and characterize 3D cameras uniformly. Desirable would be a method, which is able to measure all types of cameras equally. With this work, we want to contribute to the standardization of 3D cameras. In this case, we use a PMD camera for outdoor applications with relatively large pixels. It is shown how to determine the spatial resolution of a PMD camera from both, the amplitude and the distance image. Further, a novel method is presented how to determine the resolution enhancement in an image via gradient image evaluation. Finally, a method is proposed which evaluates the quality of resolution enhancement, when no ground truth data is available. Both are particularly interesting for the use of super-resolution (SR) applications.