| Abstract: | The evolution of digital photogrammetry from analytical photogrammetry is outlined, partly by considering the methods employed, but with the main emphasis on mathematical and physical principles. A fundamental distinction is obvious from the nature of the primary information measurement data: the image co-ordinates with analytical photogrammetry and the digital image grey values with digital photogrammetry. For both techniques, appropriate Gauss-Markov evaluation procedures are compared. The relevant relationships between image space data and object space models are presented; radiometric considerations play a more important role than before. The evaluation of digital image grey value data is no longer based on digital image correlation. Instead, a new principle of image inversion has been introduced by directly back-projecting grey values of an image on to object space models. The numerical procedures of digital photogrammetry need some stabilisation by appropriate regularisation elements. The new concept of digital photogrammetry may be applied to imagery from almost any sensor. |
