The 
The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part B3b. Beijing 2008 
There are two assumptions to construct the spatial road 
network. 
a) There is a traffic pattern in the centre of local cluster 
b) There is no significant change of speed on each road 
segment linked by nodes 
In the real traffic system, these two assumptions are always easy 
to realize. Local clusters mean a significant change of speed, 
because of high FCD density. Between every two clusters and 
along a road, vehicle can run at a relatively stable speed. 
Otherwise, there must be another cluster which should be 
detected. Therefore, local clusters can be regarded as nodes of 
road segments. 
4.2 Strategy for candidate road segment selection 
After finding nodes of road segments and projecting these nodes 
by simple cylindrical projection method with a WGS84 datum, 
each two nodes can be linked as our candidate road segments 
and in the meantime, these road segments and multi spectral 
remote sensing imagery are registered into the same coordinate 
system. After a pre-processed procedure for multi spectral 
remote sensing image, the result is used to determine the final 
spatial road network. This can be explained by figure 3. 
In figure 3, 7 nodes are first found with method introduced in 
section 3. They are a, b, c, d, e, f and g. Then each two nodes 
can be linked and 21 candidate road segments are produced. Rl, 
R2 and R3 are the road areas which are extracted from multi 
spectral image. In figure 3, this is an ideal result, but actually 
the result from pre-processed multi spectral image may contain 
noise and uncertainty, which are endurable because, from figure 
3, if the areas cover all roads in the image can be roughly 
extracted, the spatial road network can be decided by 
calculating the probability that each candidate road segment 
falls in road areas and with a given significance level. 
5. A CASE STUDY 
To illustrate the methodology proposed in this paper, a case 
study is carried out. In the area between 22°31'48.00"N and 
22°32'24.72"N in latitude, 113°54'34.86"E and 113°55T3.56"E 
in longitude (Simple Cylindrical projection with a WGS84 
There 
In figure 4, there are three inputs for the multiple testing. They 
are FCD, weight matrix and critical value. The critical value is 
obviously from Monte Carlo simulation process, which is 
introduced in section 3. After the multiple testing, road segment 
nodes are detected and then candidate road segments can be 
produced. The selection procedure is discussed in section 4. 
Finally, the spatial road network is constructed. 
5.2 The weight matrix 
Weight matrix is one of three inputs for multiple testing. 
Actually, weight is essentially important for the multiple 
testing. Figure 5 shows a comer of FCD.