he 
he 
he 
i- 
oh 
SS 
he 
int 
m 
d. 
he 
minimum and the maximum of the matching points: the 
registration process terminates if the matches exceed the 
maximum, while the process continues if the correct matches 
have not fulfill the minimum requirements. 
4. EXPERIMENT RESULTS AND ANALYSIS 
4.1 Experiment results 
To ensure the versatility and effectiveness ofthe proposed 
algorithm, 6 pairs of CE-1 images, numbered 0192, 0232, 
0397, 0488, 0522, 0621, were selected randomly in this 
research. Each pair contains both B and F view images. The 
first two pairs have the size of 512 35672 while the other four 
pairs are in the form of 512 X 36360. Based on the original SFIT 
algorithm and the proposed parallel and adaptive uniform- 
distributed registration method, 4 sets of experiments are 
conducted on the randomly selected images in this research. In 
the experimental process, the overlapping area is partitioned 
into several blocks first, then the parallel registration is carried 
out among the generated units based on the one-threaded, dual- 
threaded, tri-threaded and four-threaded process separately. The 
results of the time and speedup sheet of six groups images’ 
parallel and the uniform-distributed registration experiment is 
presented in Table 1, and Figure 5 shows the statistic result of 
the speedup of the parallel registration. 
Table 1 Time and speedup sheet of images parallel and the 
uniform-distributed registration experiment 
  
Image 0192 Image 0232 Image 0397 
  
Algorithm Thread Time Spee Time Spee Time Spee 
(s) dup (s) dup (s) dup 
  
* 
22 ? 
Standard 2260 1.00 2255 1.00 | 2308 1.00 
  
  
  
  
1 
ane 2 1481 153 1487 ^ 1352 1519 s 
n 3 1208 1.88 1218 1.85: 4i 186 
4 700. 7286 785 287 78 204 
Proposed I 948 1.00 1263 100 1276 100 
algorithm 2 714... 1.533; | 978." 129... 2052 134 
(matching 3 63 154 860 14 835 IS 
oints 215) 4 48 194 6 186 672 190 
Proposed 1 953 1490 12724 100 1988 1.00 
algorithm 2 729 31 2 1985 1.29 29051 135 
(matching 3 621 153 4868 147 8271 156 
pomise 491 194 65 186 44 191 
[15,30]) 4 
Proposed 1 67. 100. .1587. 100 1368. io 
algorithm 2 883.0. 1.32 (14216: 131.7 4197 1. 135 
(matching 3 773 1.31 1046 1.52 994 1.57 
oints=20) 4 638 183 824. 195 793 197 
  
  
  
Image 0488 Image 0522 Image 0621 
  
Algorithm Thread Time Spee Time Spee Time Spee 
(s) dup (s) dup (s) dup 
  
  
  
  
  
E 1* 2345 100 2310 100 2345 100 
SET 2 1581: 3453/45320 452. 1545. 52 
T 3 1253.41.87 1278.  L88 1251 ‘187 
algorithm 4 811. 289 2794 291. 814. 2.88 
Proposed 1 1280 1.00 1341 1.00 1145 1.00 
algorithm 2 980 129 1040 129 $76 131 
(matching 3 833 154 "4000 449; 783 1.46 
points >15) 4 64 190° 725 18 612 187 
Proposed 1 1290 1.00 = 1352 1.00 1160 1.00 
algorithm 2 995 130. 1053 1.28 898 1.29 
(matching 3 840. 154 006, 149 785... 14% 
pointse[15, 683 139 9335 184 es 192 
30]) 4 
Proposed I 1619 1.00 1659 1.00 1436 1.00 
algorithm 2 1217. 133 1974. 130. 1070 … 1.33 
(matching 3 1000 157 1080 154 044 152 
points=20) 4 225 106 853 194 6 19% 
  
  
* indicates the standard SIFT algorithm 
4.2 Results Analysis 
In this section, the performance of a parallel and adaptive 
uniform-distributed registration method was presented and 
compared with the standard SIFT algorithm by using CE-1 
lunar imageries as the experimental data. The experimental 
results are concluded as the followings: 
1) By introducing the parallel processing and the uniform- 
distributed registration as two improvement factors, the 
efficiency of the CE-lunar imageries registration was 
dramatically improved. Since the goal of the proposed method 
in this research is to acquire satisfied number of the matching 
points rather than the quantity, and in the most circumstances 
the SIFT descriptor generating operation is only conducted on 
the Gaussian octaves and DoGs at local scale, the consumed 
time for the image registration is significantly reduced. 
Compared to the standard SIFT algorithm the uniform- 
distributed method could reduce the operation time by 29.3%- 
79.2%, the average of which is 56.2%. On the other hand, the 
parallel processing is also proved to be conductive to improve 
the registration efficiency. As can be seen in Table 1, the 
speedup of 6 pairs of images could reach from 1.31 to 2.89 
under 2-4 threaded operation. 
2) Under the condition of the same number of threads , the 
processing time for the proposed method is much less than the 
standard SIFT algorithm for CE-1 lunar imageries. However, 
the improvement extent of the speedup indicator for the paper 
proposed method is not as dramatic as the standard SIFT with 
threads number increases. This because the extra calculation for 
the matching points counting and control is introduced in the 
proposed method. In this experiment, the speedup ration of each 
parallel algorithm does not growth linearly with the number of 
threads increases, this result accords with the general regulation 
of the parallel computing. 
3) By applying the parallel and adaptive uniform-distributed 
registration method, the matching points in each divided blocks 
will be extracted based on the required amount, and this 
operation terminates when the amount is reached. For the blocks 
in which the matching points could not achieve the required 
amount, the Gauss pyramid and the DOG images will then be 
introduced for the image registration, thus realizing the uniform 
distribution of the matching points across the whole image. The 
analysis results suggest that the matching points are more 
uniform distributed in condition of the maximum value is set. 
Moreover, setting the maximum and minimum as the same 
value could also improve the uniform distribution of matching 
points effectively. 
467