Traffic Spatial Measures and Interpretation of Road Network Using Aerial Remotely Sensed Data

The most common measurement of usage in road network is vehicular traffic flow, which is defined as the number of vehicles passing a specific location along a road during a unit of time. Flow is often expressed as AADT (the annual average daily traffic). Traditionally, the traffic vehicle flow is measured by using ground-based sensors, most common of which are inductive loop detectors and pneumatic tube detectors. The sampling rate of a detector depends on the application and the operating agency, typically ranging from under one minute to a full day. This paper presents methods for spatial measures and interpretation of roadway network from aerial remotely sensed image. First, this paper present the interpretation of roadway network using co-registration between up-to-date aerial imagery and geospatial database, which is called image-to-GIS co-registration. The up-to-date aerial imagery and their extracted road network features are thought as reference data, and the early GIS data are co-registered onto aerial images after orthorectified in order to interpret the remotely sensed image; (2) the vehicles are detected based on the co-registered data; (3) Based on the co-registered data, and recognized vehicle, the measurement including vehicle density, headway are conducted, further to compute the flow rate; (4) the experimental results are compared with the results from the ground-based detection. Finally, some conclusions are drawn up.