582 
(a) Pre-Processed imagery 
(b) Final spatial road network 
Figure 9. Candidate road segment selection 
6. CONCLUSION 
In the latest decade, GIS, especially WebGIS has been 
developing prospectively. GIS-T, as an important part of GIS, is 
discussed more and more. Spatial road network, the 
fundamental element, has become one of key topics. As the 
introduction of new form data, FCD, into GIS-T, there comes a 
new trend which leads to build a more dynamical and more 
intelligent traffic information system. In this paper, integrated 
the advantages of FCD and high spatial resolution imagery, a 
methodology to construct spatial road network automatically is 
proposed. 
In this paper, a new statistic is defined to describe the local 
cluster based on the character of road intersections. To obtain 
the critical value of the statistic, Monte Carlo simulation 
process is employed. Bonferroni adjustment is suggested to 
keep the experimentwise error rate to a specified level. In the 
case study, kernel density analysis is carried out to get the key 
parameter for building the weight matrix. After all the road 
segment nodes are detected, candidate road segments are 
produced. Assisted with the pre-processed high spatial 
resolution imagery, the spatial road network is finally decided. 
Besides the final spatial road network obtained with this 
methodology, it should be noticed that all the road nodes are 
detected based on FCD and candidate road segments are filtered 
based on multi spectral remote sensing imagery. Therefore, on 
the one hand, these nodes are the most compatible with FCD 
and on the other hand, these nodes match multi spectral remote 
sensing imagery very well. Thus the significance of this