Spatial Accuracy Analysis Of Ultra-short Baselines And Multiple Overlapping Images

At present,digital photogrammetry plays an important in the collections of detailed and accurate three-dimension information which is required by vast applications such as smart city,digital twin,bathymetry,industrial object measurement,environmental monitoring and so on.More and more non-metric cameras are used routinely many of those photogrammetric applications.Due to the distortions on non-metric images and traditional stereo image techniques used,low spatial accuracy is very common among those applications.In this paper,a detailed study is conducted to explore the idea of using ultra-short baselines and multiple overlapping images(USBMOI)to improve the spatial triangulation accuracy.This study includes the theoretical analysis of USBMOI spatial triangulation,a series of USBMOI numerical simulation experiments,A finding is revealed that the optimal spatial accuracy(one order of magnitude improvement)could be achieved when approximately 60 USBMOI are used comparing to the traditional stereo-image case.This finding is further proved under three real scenarios: 1)an indoor controlled close-range application using 60 USBMOI;2)a two-media photogrammetric application for underwater measurement using 65 USBMOI and;3)an underwater bathymetric application using 43 USBMOI.This paper also explores the occlusion problem in urban environments.Preliminary simulation results suggest that the occlusion percentage can be reduced through the use of USBMOI and the spatial accuracy is also improved in the meantime.It is the author's hope that the conclusions derived in this study could be used as guidelines when ultra-short baselines are employed in order to improve the spatial triangulation accuracy.