compensation, and the correSponding process comes to a Stop. 
3e A posteriori compensation of systematic errors 
After T-testing had been performed, if T>1, it denotes 
that sisnificant systematic errors exist, and a posteriori 
compensation is needed to be done. Values of a posteriori 
compensation of systematic errors for coordinates of all 
model points based on corrections of svstenatic errors at all 
tie points are obtained by using method of interpolation 
(in this experiment both method of polynomial and weighted 
mean values interpolation are used). With coordinates of 
model points after adding the compensation values, next whole 
block adjustment begins to carry out. 
Formula /2/ used for polynomial interpolation is given as: 
  
VU, =0 HA x Ha, xy a,01 aux y! +a,x"—a,x!y 
Uy T bo byx t buy baxY t DAY + bsx¥* |+ box + byx?y 
Va mOed Cx d C4 CX T 04! ex y! + cx + cy 
  
  
where x,y are coordinates of model points. 
a;» by and cg are coefficients of interpolation 
When adjustment is executed by using individual model as 
unit, the preceding 6 coefficients of interpolation are used, 
When adjustment is made by double models, all 8 coefficients 
of interpolation are adopted. 
Formula /3/ used for interpolation of weizhted mean values 
is: 
Z3 (Z,D/ a 
i=1 
/3/ 
where d; refers to the distance between voint to be deter- 
mind p and tie point i , Z4; is correction of SyStematic error 
at tie point i, Zp is the correction of svstematic error at 
point p sn is the number of tie point. When des 1,d; is 
taken as 1. 
III. Experiments on a Posteriori Compensation of Systematic 
Errors 
le Materials of experiment 
For this experiment both simulated and practical data are 
- 492 - 
ree ree = _