Change detection for traffic measurement in multi-temporal TerraSAR-X SpotLight images

Abstract:
The knowledge of area wide up-to-date traffic information is a key issue for accurate traffic forecasting and service. The observation of large areas with very high resolution radar sensors provides very useful snapshots for one instant in time. Ground moving target indication (GMTI) utilizes different techniques such as along track interferometry to detect nonstationary objects. Slowly moving or stagnant objects however are very difficult to detect but are as well of particular interest since they are typical for traffic jams. We therefore apply a change detection method to the amplitude information of multi-temporal TerraSAR-X image stacks in order to disclose stopped or jammed cars on roads and to monitor fluctuations on parking lots. As an example we analyze queues of cars waiting in front of traffic lights in Las Vegas as well as significant fluctuations in capacity utilization of parking lots around the Las Vegas Convention Center (LVCC) depending on occurring events. Index Terms: Change Detection, SAR, Traffic Monitoring, TerraSAR-X