In the central business districts of many cities, a large fraction of the traffic comes from cars cruising for parking. One study estimates that as much as 30% of the traffic in downtown areas of major cities. Cruising wastes time and fuel for drivers and is also detrimental to the municipality, as it pollutes, congests traffic and degrades pedestrian environments. Parking search assistance is one of the major unsolved problems in urban mobility. Occupancy information is currently only available in parking garages and off-street parking, where parked cars are easily counted. For on-street parking, which is more preferable to drivers since it is usually cheaper and more convenient, occupancy is not so easily obtained.We are developing a vision-based smart parking system for monitoring on-street parking occupancy. A camera network monitors parking spots throughout the city and sends images to centralized servers, where the images are then processed with computer vision algorithms to determine parking availability. Real-time occupancy information can be broadcast to drivers through a website or phone app, while aggregate statistics can be provided to parking administrators. One can further integrate navigation to nearby parking, payment and enforcement functions directly into the system. Such a smart parking system can greatly reduce frustrations of finding parking and also the noise and pollution from traffic congestion.