In many sensor systems used in urban environments, the amount of data produced can be vast. To aid operators of such systems, high-level information fusion can be used for automatically analyzing the surveillance information. In this paper an anomaly detection approach for finding areas with traffic patterns that deviate from what is considered normal is evaluated. The use of such approaches could help operators in identifying areas with an increased risk for ambushes or improvised explosive devices (IEDs).