The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Voi XXXVII. Part B7. Beijing 2008 
129 
distance of the extracted roads position to the reference, and the 
optimum value is 0), false alarm rate and missing alarm rate(the 
complement sets of the completeness and the correctness 
respectively)[5].Used the indexes above, the performance of the 
selected various parameters were evaluated(table 1). From table 
1 some conclusions in section 2 were verified. Low number of 
road region pixel can lead to bad results. The large of the road 
length pixel as row 4 in table 1 can lead the increase of the leak 
rate. One pixel width and 17 pixel length as middle road region, 
at the same time the pixels of number of left region and right 
region was set 119 can get desirable result and can not get to 
the theoretical 1% false rate alarm due to the change of the 
homogeneity. The result in figure 5 is quite satisfied. 
[1] LI Su-dan, ZHANG Cui, WANG Zheng-zhi. A study of 
road network detection method for SAR images [J]. Journal of 
Astronautics, 2002, 23(1): 17-24. 
[2] JIA Cheng-li, KUANG Gang-yao. Automatic Extraction of 
Roads from Low Resolution SAR Images [J]. Journal of Image 
and Graphics, 2005, 10(10):1218-1223. 
[3] Ridha Touzi, Armand Lopes, Pierre Bousquet. A Statistical 
and Geometrical Edge Detector for SAR Images[J]. IEEE 
TRANSACTIONS ON GEOSCIENCE AND REMOTE 
SENSING, 1988, 26(6):764~773. 
! 
Data Preprocessimg 
(block the large image, 
equivalent looks processing,etc.) 
r~~ 
unsuitable 
model parameters chosen 
suitable 
} 
The judgement of homogeneous local region or not by VI 
T F } >ST 
calculate p p p p calculate(p p A,(p p A 
>T 
] 
labeling roads points 
' unsatisfied 
the performance evaluation 
j satisfied 
elimilating holes,connection,etc. 
t unsatisfied 
Results Assessment 
j satisfied 
submit 
[4] LIU Zhen-hua, YU Wen-zhen, MAO Shi-yi. Compounding 
Reconstruction for Synthetic Aperture Radar Images [J]. ACTA 
ELECTRONICA SINICA, 2004, 32(3): 363-367. 
[5] JIA Cheng-Li, ZHAO Ling-Jun, WU Qi-Chang, KUANG 
Gang-Yao. Automatic Extraction of Road Network from SAR 
Imagery Based on Genetic Algorithm [J]. Chinese Journal of 
Computers, 2007, 30(7): 1186-1197. 
[6] Birgit Wesssel, Christian Wiedemann, Heinrich Ebner. The 
Role of Context for Road Extraction from SAR Imagery [A]. In: 
Geoscience and Remote Sensing Symposium, 2003. IGARSS 
'03. Proceedings[C], 2003 IEEE International, 2003, 
6(7):4025~4027. 
[7] Stefan Hinz. A Fusion Strategy for Extraction of Urban 
Road Nets from Multiple Images [A],In: Geoscience and 
Remote Sensing Symposium, 2003. IGARSS '03. 
Proceedings[C]. 2003 IEEE International, 2003, 
2(7): 1059—1061. 
[8] Manuele Bicego, Silvio Dalfini, Gianni Vemazza et al. 
Automatic Road Extraction from Aerial Images by Probabilistic 
Contour Tracking[J].IEEE Image Processing,2003,3(9): 
585-588. 
[9] Roger Fjortoft, Armand Lopes. Optimal Edge Detection and 
Edge Localization in Complex SAR Images with Correlated 
Speckle [J]. IEEE Trans, 1999, Geosci Remote Sensing- 
37(5):2272 - 2281. 
Figure 4: the flowchart of roads extraction from SAR imagery 
5. CONCLUSIONS 
In the article, we amplified the uncertain factors in the 
automatic roads obtained from SAR Imagery. We put forward 
the hypothesis of the SAR imagery and then built up the local 
road model. From the probability distribution theory, we got the 
mathematic function of the model base on the assumption. The 
next we analyzed the parameters of the function. From the 
uncertain property, we concluded three criteria for the 
extraction. Through tests, with the objective performance 
evaluation system, use these metrics to continually adjust 
uncertain parameters in a real time manner to get a satisfactory 
result automatically ,we got one reasonable procedure for the 
roads net extraction from the SAR imagery in practice. The 
achievements can be applied in image registration, mapping and 
GIS applications etc. 
[10] XIE Feng, CHEN Ying, LIN Yi .Road extraction from RS 
imagery based on wavelet and mathematical 
morphology [J]. Computer Engineering and Applications, 2007, 
43(22):241-243. 
[11] WEN Gong-jian, WANG Run-sheng. Automatic 
Extraction of Main Roads from Aerial Remote Sensing Images 
[J], Journal of Software, 2000, 11(7): 957-964. 
[12] SHI Wen-zhong, ZHU Chang-qing, WANG Yu. Road 
Feature Extraction from Remotely Sensed Image: Review and 
Prospects [J], ACTA GEODAETICA et CARTOGRAPHIC A 
SINICA, 2001, 30(3): 257-262. 
[13] YU Changhui, MENG Lingkui, QIN Kun. Road Extraction 
in Remote Sensing Image Based on Marked Point Process [J]. 
Geomatics and Information Science of Wuhan University, 2006, 
31(1): 59-62. 
[14] HUANG Shi-qi, LIU Dai-zhi. Analysis of Some Uncertain 
Factors in Spacebome SAR Imaging and SAR Image [J]. 
REFERENCES