Critical evaluation of stereophotogrammetric methodology with emphasis on close-range applications
This dissertation represents a critical evaluation of the photogrammetric methodology in terms of the functional model, stochastic mode, numerical processing scheme, and operational system. This study then focuses on close-range applications. For the functional model, the relationship between algebraic and physical form of the perspective transformation model is fully explored, and applications are provided for both the single photo case and multi-photo one. Different functional models are studied comparatively. As an extension of the basic functional models, the conventional approach with additional parameters is investigated from both function and numerical perspectives, and other trend removal approaches are also evaluated. Finally, the recovery ability of “calibration” parameters in the conventional approach with additional parameters is studied with experiments. Concerning the stochastic model and the numerical processing scheme, a weighted constraint model, and a parametric model with additional observations are compared with a combined model. The utilization of a variance-covariance-component-estimation technique (MINQUE) for covariance matrix estimation is evaluated. The application of Box-Jenkin’s time series analysis technique, along with a general data processing scheme in the numerical realm are outlined. Photogrammetry has been applied at close-range under different forms. A conceptualized model is developed in order to bring these together within a generalized photogrammetric family. Rasterstereography, one if the methods utilizing structured light, is found to be attractive in today’s environment. An experiment with film-based raterstereography is described. The feasibility of measuring image coordinates with an enlarger-digitizer approach is fully explored, and finally, the utilization of a home video camcorder for image acquisition.