At present, the use of aerial photography has acquired a wide scope not only for small-scale mapping of the Earth's surface, but also for the creation of topographic plans on a large scale. We know, for example, the successes of Soyuzmarkstrest, whose enterprises, using aerial photography, annually produce on a large area of topographic plans on a scale of 1:5000 and 1:2000. Large-scale surveying is also used with great success in road surveys, in urban surveying, in land management.

To increase the accuracy of the heights obtained from a stereo pair, it is natural, first of all, to strive to ensure that one meter of height corresponds to the largest possible increment of longitudinal parallax at a given scale of the photograph. Or, in other words, one should strive to increase the ratio of the increment of longitudinal parallax (Δр)1 per 1 m to the segment (m)1 on the photograph corresponding to 1 m of horizontal distance on the ground. Let us designate this ratio as t. Thus, the value of the indicated ratio acquires a significant role in characterizing the accuracy of height determinations.

The resection (Pothenot problem) is a very common method for determining the position of a new point. The presence of a multiple resection always requires adjustment. The simplest and most illustrative method of adjustment in this case is the graphical method, in which, as is known, lines of position are used. Assuming the principles of this adjustment as known, the present article examines, without unnecessary detail, the error figure formed by the lines of position of the resection, without delving into the method of its construction; a particular case of the error figure is discussed. At the station, we measure directions to reference points and then adjust them. As is known, with such a preliminary adjustment, the error figure has a number of double and triple intersection points. The essence of graphical adjustment consists in adjusting the error figure, and this naturally explains the research that has been aimed at simplifying this process. Such research has led us to an important conclusion. Namely, it turned out that the point of application of the resultant of the double intersection points and the point of application of the resultant of the triple intersection points coincide with each other, and the sum of the weights of the triple points is in a simple ratio to the sum of the weights of the double points. Furthermore, from the Pothenot problem with four points, a transition is made to any number of given directions to points. By means of simple reasoning, the validity of our theorem is proved also for these cases.