Advanced Methods for the Calibration of Metric Cameras

摘要

A new analytical approach to the problem of the accurate calibration of metric cameras is developed and specific applications are reported. The method permits an indefinitely large number of frames from a given camera to be reduced simultaneously, yet efficiently, to produce common parameters of the inner cone for all frames, as well as independent elements of exterior orientation for each frame. Because control points may be exercised repeatedly in a common reduction, very large sets of well distributed residuals can be generated from a relatively small set of control points (in principle, a complete and accurate calibration could be performed from as few as three control points). From such large sets of residuals empirical functions can be derived to account for persistent, unmodelled systematic error. In addition, it becomes feasible to establish the variation in the accuracies of the radial and tangential components of plate coordinates throughout the format and thus to establish appropriate empirical weighting functions for subsequent applications. The method is of universal applicability and encompasses calibrations from aerial photographs, calibrations from stellar photographs, and calibrations from multicollimator photographs. Applied to stellar calibrations, the method leads to improved accuracies and convenience by completely doing away with conventional requirements for precise timing of the shutter and for stability of the camera throughout successive exposures. Applied to multicollimator calibrations, the method has far reaching implications concerning both the use of existing multicollimators and the design of future multicollimators. (Author)