UWB radar systems become more and more important for material penetrating and imaging applications. Many conventional UWB signal processing algorithms for image generation are commonly based on migrations techniques which are not optimal in terms of object identification capability and calculation time. Thus, a surface reconstructing imaging algorithm has been implemented and verified by real radar data. It offers two crucial advantages with respect to the conventional algorithms: an extensive reduction of the calculation time and the ability to identify targets by the shape of the object. Several examples are given demonstrating the efficiency of this approach. It is based on a preselection of the received C-scan yielding the quasi-wavefronts of the object.