Research on local visual global localization method based on out-of-view reference of spatial point association

¹ College of Computer Science, Nankai University, Tianjin 300071, China
² State Key Laboratory of Precision Measuring Technology and Instruments, Tianjin University, Tianjin 300072, China
³ College of Engineering Design and Physical Sciences, Brunel University of London, Uxbridge UB8 3PH, UK

^* Corresponding author: This email address is being protected from spambots. You need JavaScript enabled to view it.

Abstract

The global localisation of spatial points is a critical step in tasks such as object tracking, motion analysis and pose measurement. This paper addresses the critical issue of global localisation when spatial points are scattered and cannot be contained within the same field of view. It proposes a local visual global localisation method based on an out-of-view reference through spatial point association. By constructing a local measurement and localisation model using parallel binocular vision and a spatial coordinate transformation model that associates local regions with the global reference, the global localisation of spatial points inside and outside the field of view is achieved. Experimental results demonstrate that the localisation accuracy of spatial points is less than 0.1 mm in terms of distance measurement. This method is useful for cooperative multi-camera localization and multi-point measurement in large 3D spaces.