5. CONCLUSIONS 
In this paper we have presented a program which recog- 
nizes real roads. This program can integrate the spectral 
and spatial information in extraction of particular tar- 
get. It can simulate the visual interpretation to recognize 
the roads. The ratio of the lenth to width of road is 
proved to be the special information to successful dis- 
crimination of the road. 
Although there are broken fragments on the road line in 
the result picture, the accuracy by this program is much 
better than those by conventional methods. 
ACKNOWLEDGEMENT 
The author carries much thanks to Prof. Zhao Yuan- 
hong and Prof. Chen Lan for their lots of helps and 
guidences. 
REFERENCES 
1. Barry, N. H. , 1984. Multisensor data analysis of 
urban environments. Photogrammetric Engineering 
and Remote Sensing, Vol. 50, No. 10, pp. 1471— 
438 
1477. 
. Charlotte M. G. and J. R. G. Townshend, 1983. 
The use of contextual information in the classifica- 
tion of remotely sensed data. Photogrammetric Engi- 
neering and Remote Sensing, Vol. 49, No. 1, pp. 55 
—64. 
. Philip H. Swain, Howard Jay Siegel, Bradley W. 
Smith, 1980. Contextual classification of multispec- 
tral remote sensing data using a multiprocessor sys- 
tem. IEEE Transactions on Geoscience and Remote 
Sensing, Vol. GE-18, No. 2. 
. Ruzena Bajcesy , Mohamad Tavakoli, 1976. Comput- 
er recognition of roads from satellite pictures. IEEE 
Transactions on System, Man, and Cybernetics, 
Vol. SMC-6, No. 9. 
. Stephen W. Wharton, 1982. A contextual classifica- 
tion method for recognizing land use patterns in high 
resolution remotely sensed data. Pattern Recogni- 
tion, Vol. 15, No. 4, pp. 317 —324. 
Tian Lianghu, Zhao Yuanhong, Zhan Fuxiang, 
1990. Urban thematic information extraction and 
dynamic extension detection. Proceedings of the 
11th Asian Conference on Remote Sensing, Q-37- 
1-Q-37-6. 
. Tian Lianghu, 1991. Detection of the developments 
of urban area using remotely sensed data. Remote 
Sensing Application, Vol. 4, No. 1, pp. 28—35.