The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part B7. Beijing 2008 
Figure 10 is the fusion result of extracted features with function 
(6), in which the road is better enhanced. The comparison 
between it and images of Figure 5-9 shows that the fusion result 
not only decreases the interference of linear water to road, but 
also makes the broken road in the upper left, middle of the left 
and lower right continuous by taking both advantage of 
microwave radar and multi-spectral images. 
Figure 11. Time trend got from LS-FM algorithm 
The time evolution of LS-FM algorithm is shown in Figure 11. 
It can be seen that the time increases gradually. When the 
evolution passes about 4800 pixels, the time is going to be a 
very great value and the road edge is arrived. 
Figure 12. Time evolvement image got from LS-FM algorithm 
Figure 12 is the final time evolution result. Colour from blue to 
red indicates the change of the time from small to large, and the 
deep red color shows the non-road information with large time 
value. It can be seen that the LS-FM algorithm can better 
overcome the interference of non-road objects in Figure 10 and 
extract the road. But because of the existence of the objects 
similar to road, the road extracted has some irregular swells. 
Input binary time evolvement image 
Figure 13. The flow chart for road connection 
In order to get one pixel width road, the result in Figure 12 is 
processed with methods in Figure 13. 
Figure 14. The thinned time evolvement image 
Figure 15. The extracted road image