intuVision VA’s patented algorithms detect vehicles and calculate their speed with high accuracy in different environments, from highways to city streets, and to help neighborhoods and companies promote safe driving. Our speed detections provide data many options for segmenting data by vehicle type and lane, and can provide either aggregate speed metrics for roadways or incident alerts with vehicle snapshot.
intuVision VA’s classification allows the vehicle types of interest - typically we recommend 4 categories, such as: motorcycles or bikes, passenger vehicles, mid-sized vans and trucks, and large trucks/buses. These categories can be customized to specific project requirements.
For traffic studies or environmental studies, looking at general speed trends by vehicles over time, data can be aggregated across vehicle types and roadway lanes, giving an over-all idea of general speed trends over time and by vehicle type. For example, this would output the average speed per hour of semi-trucks passing in a far lane, and can give guidance as to adjusting speed limits or posting speed patrols. Typically used with this form of data collection, the intuVision Traffic Roadway Lanes feature automatically separates the speed detections by lane. The lane marks are added by simply choosing the number of lanes and clicking and dragging the template on the view. When used together with intuVision vehicle classification, by setting up one speeding vehicle event, all events are segmented by lane and by vehicle type.
For neighborhoods or businesses, looking to ensure safety in their roadways and parking lots, the notification and security tools of intuVision VA help support knowing exactly when and who is speeding. Alerts can be sent via email, reviewed at a later date in the intuVision Review, and can include not only time period, calculated speed, but also a video clip of the incident and a snapshot of the vehicle. Real-time alerts can allow security to intervene immediately, or weekly overviews of data can point to people consistently breaking the rules. Typically vehicles owners will be known in these environments vehicles will be registered, so snapshots, video, and calculated speed allows for an intervention.
Speed detection results can be reported and exported in various formats as text, graph and Comma Separated Value files as well as monitored in real-time in intuVision VA’s comprehensive Review Application. intuVision Web API allows pulling the results into another application, such as a dashboard, and intuVision HTTP triggers allow sending counts of events in 15-minute intervals.
Included in both the intuVision VA Traffic module and intuVision VA Security module , our speed detection produces accurate automated vehicle counts, without needing to install intrusive physical equipment other than a camera; while also providing a variety of secondary security events and traffic event data to enable a full traffic analysis.
Flexible intuVision VA classification training is customizable to specifics in each camera view for highest accuracy and the trained categories can be used on similar camera views.
Comprehensive output triggers and web API make it effortless to bring counts into your front-end for additional analysis.
intuVision VA Vehicle Counting is easy to set-up, yet allows full control of the processing with detailed settings for highly accurate results.
By detecting headlights and taillights in nighttime views, intuVision speed detection maintains accurate results at night. Flexible processing settings allows for daytime processing with standard intuVision object detection and tracking, and nighttime processing in vehicle (headlight/taillight) detection mode.
intuVision VA speed detection is ready to deploy with all ONVIF compliant cameras and major VMSs. The only requirement is a camera with an unobstructed view of the traffic.
High mounted cameras with close to a top-down direct view of the traffic will provide the highest accuracy as shown in the scene on the left. Accuracy decreases with lower mount cameras viewing the traffic at an angle, as large vehicles closer to the camera will obstruct the vehicles in farther lanes. Camera mounting height recommendation will depend on the number of lanes to be monitored; higher camera placement is required for wider roads and busier traffic. Highest accuracy will be seen in views where vehicles move perpendicular to the camera, as opposed to directly towards the camera. However, with sufficiently high camera placement, such as the below, speed can be calculated for vehicles moving towards the camera.
Over 90% accuracy, excellent camera view. Slightly higher or larger view would be ideal, as vehicles in the far lane might be blocked by those closer, but this happens rarely and speed incident highly unlikely if two vehicles in view.
Over 85% accuracy in center two lanes, 80% accuracy in far lanes, decent camera view. Camera could be tilted down to see more of vehicles are they are close the camera.
For most accurate vehicle detection at night, intuVision VA requires centered view of vehicles driving away from/towards the camera such that both headlights or taillights are visible for each vehicle.
intuVision VA Traffic licensing options are flexible for working with realtime streaming or recorded video to support both permanent installations and short term/periodic traffic studies.
For recorded video intuVision VA has faster-than-realtime processing capability, which allows processing video up to 20x faster than realtime, and getting results in a fraction of the time for faster delivery to your customers.
Licensing for realtime analysis is per camera while the post analysis (for recorded video) is at an hourly rate. We are able to offer solutions customized to your projects whether you are interested in short term studies or setting up permanent smart city installation.
intuVision VA speed detection is being used around the world for Smart City deployments, traffic studies, and to secure neighborhoods and parkinglots as highlighted below. Please contact us for more information on these use cases and how our solution might apply to your vehicle counting scenario.
Customer is using intuVision VA Speed Detection along with classification to estimate the amount of pollution being generated. As different types of vehicles generate different levels of pollution depending on their speed - calculated speed and classification was ingested into the customer's third-party application as raw data and included in their calculations.
In conjunction with intuVision VA Speed Detection, this customer is also using intuVision congestion detection to track periods of time when vehicles come to a stand-still. As speed calculations rely on moving vehicles, this was used to ensure only accurate speed data of standard traffic motion was used in their calculations. .
Customer had received multiple complaints about specific neighbors dangerously breaking the speed limit in their gated community, but without proof to back up the claims, nothing was able to be done. Using a short-term installation of intuVision VA and a camera mounted on the front building, needed proof was collected to ensure that the person speeding was confronted. Recording the video at a high resolution and sub-sampling the intuVision VA processing, was enough to give a clear view of vehicle's license plates.
A parking garage wanted to ensure that customers do not speed dangerously in the garage, causing potential accidents and property damage. Despite low camera angles and variable light from headlights, intuVision VA speed detection was successfully implemented. Camera views to allow not only speed detection, but also a view of the vehicle's license plate allowed intuVision email notifications with snapshots to show the estimated speed of the vehicle and their license plate. As vehicles paid to exit the parking lot, those who were detected as speeding could be approached or black-listed from future usage.