Automatic and accurate passive target centroid detection for applications in engineering geodesy

Abstract

A novel automatic and accurate passive target centroid detection approach is proposed and developed which has numerous fields of applications, such as in engineering geodesy, photogrammetry, robotics and navigation. It is robust and reliable with respect to poor environmental conditions, such as low lighting, dusty situations and skew angle targets. This research is conducted in two parts. First, an optimal target selection and its accurate, automatic and fast centroid detection are performed. Afterwards, two case studies of the proposed approach are presented for automatic target pointing by means of the telescope camera of an image assisted total station (IATS) (here Leica Nova MS50 MultiStation) and automatic target measurements for a digital camera (here Nikon D750). The results show that a circular target with a line pattern consisting of four intersected lines is an optimal pattern. The accuracy of the proposed method varies depending on the image resolution and distances to the objects. We performed experiments for distances between 1.6 and 28 m for the IATS and achieved sub-millimetre accuracy.