The conventional “nodal point method” of determining the configuration of a three-dimentional object by photograms taken with a pair of cameras leads, in general, to erroneous results for objects at short distance, even if the camera-lenses used were completely free from distortion. The reason is due to the fact that the position of the center of an out-focussed image on a photogram depends both on the location of the corresponding object point and on the position of the effective stop of the lens used, while no consideration is taken with respect to the latter in the nodal point method.
An alternative “chief ray method” is proposed which utilizes the fact that an object point which makes an out-focused image on a photogram lies on the chief ray in object space corresponding to the central point of the blurred image, with description of two methods of determining the chief ray in object space for each point on the photogram. A special feature of this method is its validity even for the case when the lens sed had a large amount of distortion such as of the fish-eye type.
A special lens is suggested for close-range photogrammetry that simplifies the execution of the chief ray method, and another that admits the use of the conventional nodal point method with no errors.
