4 
b) The triangulation of the fictitious strip with the grids has been executed with the same 
operating procedures that, as we will see later on, have also been used for the triangulation of 
real photograms, namely: focal length 100.35, size I 4° X 140 mm 2 , overlapping 60 %, model 
scale 1 : 10.000, ratio Z/f = 1.2. The absolute orientation of the first model has been performed 
according to the same procedure previously described for the grid model. The grid corresponding 
to the camera to be bridged was slightly moved before starting each orientation: the routine of 
interior orientation was then repeated all over. For the relative orientation of the models we 
have used the parameters of only one camera, and have measured the Y parallaxes on the six points 
in standard position; the scale transfer has been obtained, as usual, by means of b x variations of 
the camera to be bridged until the height of the nadiral point had reached the previously registered 
value. The instrumental coordinates of the six standard points of each model and the orientation 
elements have then been registered. The instrumental coordinates of the points are automatically 
determined in the coordinate system of the first model; therefore, it is not necessary to perform 
on them any transfer computation. 
We have executed two independent bridgings of 21 photograms each. The table of the diffe 
rences of the corresponding coordinates of the points common to two consecutive models does 
not show any special interest and, therefore, is not included in the present text. The means M 
and the mean square values A/M 2 — VE A z / n (m. s. v.) of the differences, independently 
computed for the three points of the nadiral line of each model [a , N , b) and for each coordinate, 
are listed in table no. 2. 
The general negligible value of the means M leads to the conclusion that no remarkable syste 
matic behaviour is present along the strip. 
Figures 3 and 4 reproduce the behaviour of the orientation elements in the subsequent models, 
in the first and second measure. We have drawn the graph of b x next to that oi A cp and the graph 
of b y next to A oj . The correlation between the orientation elements that have been coupled is 
evident both in the first and second measure. The saw pattern of A co and b y , so evident in the 
first measure, is probably due to the different behaviour of the instrument in the « base in » and 
« base out» situation; it is, however, less regular in the second measure. In both strips b z regularly 
decreases and reaches nearly the same value in each of them. In figures 3 and 4 the An is not 
shown among the orientation elements: A x is in fact arbitrarily variable, due to the fact that the 
initial position of each grid was changed from time to time. 
Figures 5 and 6 represent the planimetric and altimetric behaviour of the subsequent models 
for measure I and II. The planimetry is definitely different in the two measures. In the results 
of the first measure we can see a noticeable and regular increase of the scale, which is not so large 
in the second measure. On the contrary, the latter shows a general curvature that does not appear 
in the former. The altimetric behaviour is similar in both cases: the two strips denounce a pro 
gressive increase in height, more regular, but also larger, in the second measure, and a screwing 
deformation of the strip itself which causes in the last model a difference (at the model scale) of 
90 /,tm in measure I and of 66 in measure II between the average height of points 3, 4 and the 
average height of points 5, 6. 
The data we obtained from the two above described groups of experiments let us conclude 
that the behaviour of the instrument is quite satisfactory. We can reach this conclusion through 
both a qualitative and quantitative analysis of the results. As far as the global precision in the 
plotting of a model is concerned, the values of the mean square deviations, contained in table no. 1, 
are, with no doubt, to be considered good, comparable with those that can be obtained by means 
of a precision plotting apparatus. All the more if we think that these results contain not only 
the instrumental errors of mechanical and optical origin, but also the errors of absolute orientation 
of the model. Moreover, from the study of the shape of the 6 plotted models, we can evaluate, 
though in an approximative way, the angular errors of relative orientation due to the accidental 
errors in measuring coordinates. Usually, these models present an error in d co and 6 x , whose 
entity is, on the average, smaller than 20". Such a behaviour being almost negligible but 
systematic, we think that it may be ascribed to mechanical causes.