The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part Bl. Beijing 2008
3. THE VARIATION OF MEAN REDIUS OF
CURVATURE AND THE INFLUENCE ON HEIGHT
SURVEY
We choose a scene of ERS-1/2 satellite image data of HerBei
ShangYi areaed about 50Km* 50 Km . We measure the
coordinates of longitude and latitude on the maps, and caculate
the mean redius of curvature of earth with Krassovsky ellipsoid
parameters by 1954 BeiJing coordinate system ,and then gain
the variety of the earth’s mean redius of curvature.
The coordinates of longitude and latitude of the area is as
follows:
Picture 2 : the selected area’s coordinates of longitude and
latitude
It can be seen from picture 3 that the variety of earth’s mean
redius of curvature with change of latitude :
While the latitude adds, the changes of the earth’s mean redius
of curvature is linear. When the latitude adds 0.1 degree, the
earth’s mean redius of curvature adds nearly 100 meters . So the
earth’s mean redius of curvature changes very much along with
latitude. How much influence is brought on spomebome InSAR
height survey by so much change? Then it is analyzed
quantificationally in detail.
Base on InSAR height survey principle, the formula is as
follows:
R? + (R e + H) 2 - 2cos6K,(^ +H) = (R e + hf (1-1)
Caculate the differential coefficient of h to R e ,
dh _ H-2cos9R l -h
~dR e ~ h + R e
We think Re has errors, and the rest parameters are true.
From the law of propagation of mean square errors ,we can get
the expression:
While 9 = 20.8° > B = 354.56m( the baseline parameter of
ERS-1/2 ) , h = 1400m (the area’s mean ground height) , Re =
6374732.39m (the local earth’s mean redius of curvature ) ,
when m Re = 1 m, you can get m h - ±0.87 m. So 1 meter’s error
of mean redius of curvature can bring the mean square error of
0.87m . Thus it can be seen, ignoring the variation of earth’s
mean redius of curvature will cause quite much the height
survey errors. It is required that on InSAR height survey, it
need to require that the corresponding ellipsoid radius to the
different image coordinates is different too.
Then it proves whether we can take the earth’s surface as
spherical surface on InSAR Height surcey in selected area .
From picture 4, it can be seen that in that way it will cause
much more errors, and the errors distribut long the direction of
latitude. So it is not feasible to take the earth’s surface as
spherical surface in that area coverage .
Picture 4 DEM comparisons true DEM with DEM taken
ellipsoidal radius as constant.
Picture 3 : Variation of earth’s mean redius of curvature with
change of latitude