The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part B4. Beijing 2008 
respectively, the comparison to the result of road updating by 
using “revision based on change detection” method and current 
used method is presented. 
The change detection experiment indicates, about 98% of the 
disappeared or partial changed roads is detected successfully on 
average (check-out-ratio), about 55% judgements of whether 
roads are changed or not is right on average (correct-ratio). 
Although there are some wrong judged roads, the change 
detection method was able to detect a majority of disappeared 
or partial changed roads. 
TEST NO. 
METHOD 
TIME 
OMISSION 
1 
A 
4'18" 
2 
B 
8'24" 
0 
2 
A 
7' 07" 
0 
B 
6'45" 
0 
3 
A 
4'08" 
2 
B 
8'59" 
0 
4 
A 
4'05" 
1 
B 
6T8" 
0 
5 
A 
3'48" 
0 
B 
5'07" 
0 
Table. 1 Comparison of the current method and revision based 
on change detection method on map road updating 
Note: in “method” column, “A” represents current method, “B” 
represents revision based on change detection method. 
The comparison of our “revision based on change detection” 
updating method results with the results’ of currently used 
updating method indicates: 
1) On the time efficiency of map road feature updating, in 
testl, 3, 4, 5, revision based on change detection method is 
lower than current updating method, because there are omission 
roads exist by current method, and the updating workload is 
less than revision based on change detection method. In test 2, 
the workload are the same to current method and revision based 
on change detection method, time costing are the same. 2 
2) Among 5 map sheets, there are 3 map sheets exist 
omission of disappeared or partial changed roads using current 
updating method, while there is no omission roads existed 
using the updating method this paper presented. 
4. CONCLUSIONS 
This paper put forward a method to update map road feature 
named “revision based on change detection” method, it divided 
the updating to change detection and map revision. In change 
detection, this paper put forward an automatic method to detect 
the partial changed or diminished road, and a semi-automatic 
method to detect the newly added road. 
Experiments indicate: 
1) The change detection methods this paper put forward 
are able to detect disappeared or partial changed roads 
and the newly added roads efficiently, improves the 
level of automation of change detection. 
2) The road map updating methods of “revision based on 
change detection” this paper put forwards avoids 
omission of disappeared or partial changed roads, and 
improves the automation level of road map updating. 
Further research is needed on taking knowledge such as surface 
features’ height, road joint relation into account to improve the 
detecting-out factor and judgeing-right factor in road change 
detection. More experiments are needed to testify the feasibility 
and stability of the updating method this paper put forward. 
REFERENCES 
Zhang Li, Zhang Zuxun, Zhang Jianqing. The application of 
Wallis filtering to image matching. Journal of Wuhan 
Technique University of Surveying and Mapping, 1999.3,24(1). 
Ivan Laptev, Helmut Mayer, Tony Lindeberg, Wolfgang 
Eckstein etc. Automatic extraction of roads from aerial images 
based on scale-space and snakes. Technical report 
CVAP240,ISRN KTH/NA/P-00/06-SE,March 2000. 
Michael Kass, Andrew witkin and Demetri Terzopoulos. 
Snakes: active Contour models. In: Brady I M, Rosenfield A 
eds. Proceedings of the 1st International Conference on 
Computer Vision. London: IEEE Computer Society 
Press, 1987.259-268. 
Gruen, A., P. Agouris, H.H Li. Linear feature extraction with 
dynamic programming and globally enforced least squares 
matching. Automatic Extraction of Man-made Objects from 
Aerial and Space Images(A. Gruen, O. Kuebler and P. Agouris, 
editors ), Birkhauser, Basel, 1995, pp.83-94. 
David J. Kruglinski. Programming Visual C++6.0. Beijing 
Hope Electronic Press. 2003. 
1222