Photogrammetry is a flexible, accurate and non-contact measurement tool, which has been proven to be very profitable for many industrial applications. In order to achieve robustness and full automation of the measurement process, all commercial photogrammetric systems use physical retro-reflective markers. However, attaching markers to big structures is often impractical. Recently, big interest emerged concerning the use of natural markers, such as lines and conics, in industrial photogrammetric processes. In this paper we present a novel approach for line detection and matching aimed at achieving good performance with industrial images acquired under varying illumination conditions.
展开▼
