EN =
COMMISSION | (11 $ |
A ppencli I5 1 INIT
i I ] i A Dr 343 r- - |
i e IN A JIAZIC OF |
L
IE vr VO A, 1956 | 1
Use of Infrared Oblique Photographs in Aerial Triangulation
by U. V. HELAVA,
Photogrammetric Research Division of Applied Physics, National Research Couneil
of Canada, Ottawa.
Assuming the terrain to be a plane, and an aerial photograph to be an exact central
mjection of the terrain, there exists a collinear relationship between the plane of the
main and the plane of the photographic image. To a given straight line in one plane
amesponds a definite straight line in the other one. This relation can be used in an
luempt to determine from oblique infrared photographs a straight line on the ground,
lnorder to control the transversal bend of an aerial triangulation strip. A series of
ilie photographs made in the direction of the flight line of the triangulation strip and
weing the total distance of bridging are needed for this purpose. In selecting the
!ulination angle the possibility of using the image of the horizon for adjustment of cross
it should be considered.
The basic idea originated with Mr. P. E. Palmer, formerly Chief Topographic Engi-
gr of the Topographical Survey of Canada, who submitted it to the Photogrammetric
search Group of the National Research Council of Canada for study and experiment.
Je present paper is a short description of the results of this investigation. A more
iailed publication is under preparation.
feoretical considerations.
The collinear relationship is only a hypothetical basis and starting point of the
uhod developed. In reality, the terrain is not a plane and the projection is not an
"mé central projection due to lens distortion and atmospheric refraction. A straight line
1 the oblique photograph will generally correspond to an irregular line on the earth
uface. The exact shape of this line, or at least the relative position of a limited number
{selected points of the line, should be first established. In addition, when the conctruc-
E of a very long line is required, the effect of the particular map projection must also
&taken into account.
In further treatment of the problem an intermediate cylindrical map projection is
med. Relative positions of ,straight line" points are determined in this projection;
‘etransformation of these points to the final map projection is then considered. In
ymection with the cylindrical projection the further assumption is made that the refe-
Jue cylinder is tangential to the earth’s surface along a great circle passing through
ith the end points of the bridged strip. The other properties of the projection are of
ior importance, because the width of the strip of aerial photographs, and accordingly
width of the projection zone which will be used, is so small that there are practically
? differences between Gauss-Conformal, Cassini-Soldner or Central Cylindrical projec-
“ns, For simplicity the last mentioned projection is accepted. In order to introduce a
indinate system into this projection let us call the great circle of tangency the X axis;
Mrdinates will be measured perpendicular to it.
Let us consider a single oblique photograph. Assuming a distortion free lens, a
"ightline on an oblique photograph together with the center of projection defines a
{me in the object space. (x in fig. 1). All the points located in this plane lie on the
ight line in the image plane. In the general case the plane x is an oblique plane. If
{ever it includes the nadir point of the oblique photograph, it is a vertical plane and
le intersection figure in the map is identical to a great circle. In this special case the
iiblem is easy to solve since only the corrections due to lens distortion must be introduced.
le position of the line ean also be selected in such a way that the nadir points of two
ilique photographs are included. Such a line may cover quite an appreciable distance,
ud Could be successfully used for control of the transversal curvature of shorter
T'ilgings,
—
i