A close-range photogrammetric system for mapping smooth nonplanar objects using shape continuity constraints and nonmetric cameras.

摘要

An adjustment system is presented which produces improved coordinate accuracy over conventional bundle adjustment techniques. The system integrates two main components: (1) simultaneous bundle adjustment and (2) surface modelling with bicubic smoothing splines. These two components interact to sequentially eliminate image coordinate errors caused primarily by film deformation. Unlike methods incorporating geometric constraints with the bundle, this system does not define or constrain the mathematical shape of the mapping object, but only assumes it is a smooth, continuous surface of an arbitrary form.; Several experimental tests using simulated and measured data, from both semi-metric and non-metric cameras, were performed to empirically evaluate the concept of this system. Results from these tests were analyzed using statistical factorial design and regression methods. The variables and their configurations which affect the system are discussed. Conclusions, future research suggestions and general applications with the system are presented.