[COMMISSION f£ YI1
Appendi»
PHOTOGRAMMETRY, Vol. XlE4, 1956] 1
V
; uS UR.
I P A CRIES IF
the IN! UR YY OJ
The S.D.L, A New Method for Stereoscopic Measurements
and Plotting
by H. YZERMAN, Woschnau, Switzerland.
Stereoscopic observation is an integral part in Several photogrammetric procedures,
cv of this delicate attribute of human effort is often overestimated and con-
The accuracy pointing. Yet, if met with proper tools it can be still surprising. The Stereo-
sequently er an attempt to realize these tools and to reduce the error of stereoscopic
Direct-Invers 18 height to less than half
ó: its value.
It is known, that
the interchange of
ortho- and pseudo-
stereoscopies improves
the parallax measur-
ements.
The following er-
rors can be eliminated:
Firstly those of Fix-
ation disparity (1)
and secondly those
caused by the Fertsch-
effect (2).
See figure 1 where
0,0,P be an epipolar
plane, and 0, the pro-
jection centre 1 or the
left eye, and 0, the
projection centre 2
the right eye, 0, the stereo-centre of perspective. P is a point of the orthoscopic profile.
lis the floating mark situated in the plane of reference R.
Two corresponding rays 0,P and 0,P intersect the reference plane in respectively P,
id P. If the points P, and P, are observed from respectively 0, and 0,, then the rays 0,P;
and 0,P, define the point P’ of the pseudoscopic profile, which is shown by the dashed line.
Vesee that 0,0, P,P, are points of a complete quadrangle of which the two sides 0,0, and
| ?P, are parallel. We know therefore that the cross ratio 0,0,0 Eno = —1, and thus that
I = —00,. This means that the interchange of stereoscopies results in inversion of the
model in the direction of 0..
For à small parallax P,P, of point P the quadrangle PP,P'P, is a parallelogram where
IM equals MP".
This means that the interchange of stereoscopies results in equal inversion of the
del with respect to the reference plane.
Or in other words: If
LM with the model point
lake the |
Fig. 1.
; caused by Fixation disparity or Fertsch-effect, the parallax
P seems to be nil, then the interchange of stereoscopies would
atent parallax visible in the order of twice its value. This could not only be
Oed with an increase of B/H ratio or magnification of two times. Still more
tallenging is, that a true zero-method is available instead ‘of an approximation method
With all its subjective aberrations.
All this is not new
ad . It is surprising therefore that the interchange of the stereoscopies
55 still stressed (
3) in order to obtain more reliable results, and thus higher efficiency.
1
|
T
i
