The Speed Factor – Video-Based Car Counting

Camera affordability, easy access to data signals and increasing accuracy of intelligent video algorithms have made video-based counting solutions more affordable and more in demand for companies and municipalities that want to better manage traffic flow.  When considering video based counting, a question that often arises is at what speed can such systems count? Video-Based Car Counting – Speeding Vehicles

When answering this question, we must first clarify the problem.  There are a wide range of intelligent video solutions that have specific characteristics and target markets.  This is also the case with counting.  Some solutions are targeted for counting people, counting vehicles on the highway, determining groups of people or things, etc.  In this case, let’s consider counting used for parking management situations, which can include garages, outdoor lots, access roads.  For this market owner/operators are looking for a low cost, accurate system which typically counts at a transition point, such as a gate, an entrance/exit or an access road location.  In most cases, vehicles in these areas are moving slow, but there are two speed scenarios to be considered; high speeds (in a low speed zone) and stopped vehicles.

FAST MOVING VEHICLES

The first scenario, high speeds in a low speed zone, is important for two reasons.  Drivers are sometimes in a rush and may be a little heavy on the accelerator at times.  For these instances, you don’t want your counting system to not detect the vehicle due to its speed, and hence miss the count.  A similar request, is the ability to measure the speed of the vehicle and alarm on a user defined threshold.  Both scenarios are accommodated in a typical video based counting system. 

The Speed Factor in Video-Based Car Counting

This video shows parking lot counting at increasing speeds.

This video shows 11 parking lot counting trails.  In the first passes, the vehicle’s speed is approximately 5 mph.  Successive trials increase the vehicle’s speed up to 25 mph.  This particular system will likely count at much higher speeds, but these scenarios were too dangerous within the testing parking lot.

It is important to note, that there are systems that can count vehicles at very high speeds (100 mph+), however, these systems are specialized for highway type installations and are typically not suited for a parking management type scenario.

THE STOPPED VEHICLE

The other scenario that often arises in parking management situations is the stopped vehicle.  This typically occurs when counting at an exit, where a vehicle is attempting to leave a parking area, but must yield to traffic for an extended period of time.  This situation can also occur during event parking, where traffic is stalled during highly congested times, such as the conclusion of the event, or the end of a work shift.

Car Counting – Handling Stopped Vehicles

Counting algorithms use several techniques to avoid counting stopped vehicles or very slow-moving vehicles multiple times.

In these cases, a poor or marginal counting solution may count the same vehicle several times as it enters the counting view, stops, and then moves again.  These counting errors can be significant, especially during times of vehicle congestion.  Accuracy, especially when counts are used for a parking management type application, is very important.  Miscounts can result is poor customer satisfaction, as well as, safety concerns when patrons are sent to an area to park and there are no available spaces.

Today’s counting algorithms use several techniques to avoid counting stopped or very slow-moving vehicles multiple times.  One method uses the empty background image of the scene to understand when a vehicle is not present.  If the background has not become visible, the current vehicle has not passed through the scene, and only one count is logged.  Even if the vehicle remains in the counting zone for an hour or more, the system will only log a single count.  This approach can recognize the empty background even if the view consists of only a small strip, perhaps only a few inches.  This is not an unusual condition, as congested vehicles waiting to enter or exit an area traditional leave very small gaps between their vehicle and the one directly in front of them.

CONCLUSION

Accuracy is important when using counts to management your traffic flow or parking operations.  When considering the use of video, or any technology, to provide counts, ensure you understand its capability to handle issues around vehicle speed, including fast moving vehicles and fully stopped ones. Doing so will ensure you have accurate counts for real time management and long-term traffic flow improvements.

Want to learn more about using video for counting?  Check out videos and product capability on our Counting Solutions web page.

This entry was posted in Video Analytics and tagged , , , , , , . Bookmark the permalink.