The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part B5. Beijing 2008 
607 
value and the minimal value in the direction of X> Y and Z, 
and the number of cubes is decided by equations(8). 
( X — X \ 
m = INT I —— shl + i 
( Y -Y ^ 
n = INT max min 
l t ) 
/ = /Ai7^- max ~ Zm ' n l + l 
+ 1 
(8) 
where m,n,l = numbers of cubes in directions of X,Y,Z 
And choose a t as the side length of the special cube. Finally, a 
relation between the side length and the resolution is gained 
through comparison and analysis by Figure 2 and Figure 3. 
corresponding points among these stations can be selected 
quickly and accurately with this method. Furthermore, the 
errors brought forth by the data which are not exact integrity in 
each basic cube are decrescence. 
Figure 2 Side Length and Calculate Time 
Figure 4 Search the Corresponding Points 
3. Registration with weight values 
The distance between the original station and a point in this 
station is very important. Because the point valid or not is 
directly impacted by this distance. So weight value which is 
decided by that distance of each pair of corresponding points is 
considered when calculating the parameters for registration. 
Thereby each point which belongs to different cubes in 
different stations can be registered with the optimal parameters. 
If there are n pairs of corresponding points, the distance 
between this point and its original scan station is 1. 
_ 0.0018 
-a ^ 0. 00175 
g .2 0. 0017 
§ « 0.00165 
“ ’> 0. 0016 
^ 0.00155 
CO LO CO LO t*— 
. CD ... 
O »-H —' -h 
O 
side length ofspatial cube(m) 
n 
S= ^ / ( , and the weight value of this point P, p. = $ ~ h 
i=l 7 
the weight value matrix P can be defined as equation (9). 
. So 
p u 0 0 0 
0 /? 22 0 0 
0 0 ... 0 
0 0 0 p„ 
Figure 3 Side Length and Std 
The optimal side length should be in the range of 20 times to 25 
times of the resolution, and this value can be confirmed to make 
it possible to obtain the accurate parameters for registration 
with the minimal error > the least iterative times and the 
shortest period. 
2. Searching the corresponding points. 
The functional model of indirect adjustment can be expressed as 
follows equations: 
A = BX -l 
D = cr 0 2 Q = a 0 2 P 
(10) 
Calculate the distances between each point and the geometry 
center point in one basic spatial cube of one station, choose the 
nearest point P, and in the corresponding cube of the other state, 
find a point Q that the distance between P and Q is the shortest. 
Points P and Q are a pair of corresponding points. All the 
Because: 
V = Bx-l (li)