Local law enforcement agencies spend significant time every day managing traffic and parking violations, such as vehicles that exceed the speed limit, make illegal U-turns, drive in the wrong direction, or dwell in no-parking zones. They also deal with other infractions, ranging from trucks driving in lanes meant for passenger cars, to jaywalkers crossing streets illegally, and bicyclists riding in pedestrian-only areas. Although these violations may not always pose extreme danger, they are public safety hazards.
Police cannot monitor every roadway, sidewalk or intersection, even if their dispatch remotely monitors video surveillance (CCTV) networks. That’s why it has become vitally important for municipalities and their law enforcement agencies to complement their video surveillance systems with video intelligence software. This blog post explains the functions of comprehensive video content analytics systems that send alerts when traffic and parking violations occur, and aggregate and analyze such incidents over time.
Because video content analytics software can process and index objects that appear in video footage, the technology can be leveraged to identify both unique objects and behaviors as well as overall trends. Powered by deep learning and artificial intelligence, video content analysis enables complex object extraction, recognition, classification, and indexing activities, thus making video searchable, actionable and quantifiable. Operators can conduct fast, filtered searches of video footage across multiple cameras, based on object characteristics, size, color, direction, and speed.
Officers must often review traffic videos after a traffic accident or bottleneck has occurred. By filtering video for relevant details, such as direction of traffic, car color and type or even pedestrian identifiers, such as clothing color or gender, video content analytics operators can focus video review to pinpoint relevant evidence that supports their investigation. The ability to classify information in video over time allows video analytics users to uncover quantitative data about trends, patterns, and normative conditions by analyzing aggregated video data and benchmarking normal activity. By so doing, operators can then configure alerting logic to trigger notifications for behaviors outside the determined benchmarks Operators can set up alerts for dwell time, line-crossing, lighting changes or anomalies, as well as pedestrian or vehicle speed, counts, and direction. If patrol officers are alerted in real-time when violations occur, they can more quickly and accurately identify and respond in those situations.
Video analytics operators can set line crossing rules for a variety of scenarios, including entrances or exits to buildings, or roadways. An operator might create an alert at a particular intersection or stretch of roadway to identify illegal U-turns or detect pedestrians entering a vehicle-only area or a vehicle entering a pedestrian-only zone.
A video analytics system can be configured to alert operators when a vehicle or person dwells or lingers in a pre-marked space or the frame of a video camera for a duration that exceeds a pre-set threshold. For example, alerts can be sent for vehicles dwelling in a no-parking zone, or an area where it is not safe to live-park. The analytics system can similarly be configured to send an alert only if a vehicle is parked in a no-parking zone during a specific time period (i.e., if there is no parking allowed between 8 PM and 5 AM, or when a road is closed to traffic). Dwell alerts can prove extremely useful, for example, when a disabled vehicle is in a travel lane, or a breakdown lane, slowing down traffic or creating a hazardous situation. Video analysis can produce heatmaps that illustrate not only where people or vehicles linger, but also how long they typically dwell in an area; this can be used to indicate common traffic jams or taxi cab waiting areas.
Video intelligence heatmaps can be superimposed on video footage to indicate highly trafficked paths and vehicle or pedestrian traffic insights. The heatmaps are key to setting benchmarks for normal activity, so that alerts can be configured to detect and notify when anomalies occur. Operators can also define alerts to notify when traffic violations, such as vehicles moving against the flow of traffic or making illegal turns against expected behavior.
Using “in the wild” license plate recognition technology (LPR) that detects and identifies license plates in unconstrained environments leveraging surveillance cameras, police officers can search for and locate a traffic violator, based on a complete or even partial license plate. License plates associated with a traffic violation can even be added to a digital watchlist for alerting when the plate has been detected across a CCTV network.
LPR can also be used to identify drivers that have exceeded a speed limit; the system can calculate a vehicle’s speed by noting the time recorded when the vehicle license plate passed by a camera at a roadway entrance, and the time when it passed a camera at the roadway exit. Operators have the option to create a custom report that presents all the vehicles that exceeded the speed limit, and can then export their details so that law enforcement agencies can issue them tickets.) To gather traffic and speed statistics, video intelligence operators can also use LPR technology to calculate the average speed of all vehicles that drove on a specific road (by day, hour, etc.), as well as classify the vehicles according to type (truck, van, passenger vehicle, motorcycle, etc.) This ability to combine LPR with data analysis helps cities make smarter decisions and more effectively enforce traffic laws.
In addition to average traffic speed data, police departments need trend data to uncover where and when other traffic and parking violations typically occur. Incidents may be related to construction projects or seasonal events, or certain times of the year, time of day, etc. For example, a video analytics operator may need to know how often illegal U-turns occur at a particular roadway, or how often pedestrians “jaywalked” across a particular street. Sophisticated and comprehensive video intelligence software is able to aggregate video data over time and visually present that data in dashboard reports to make it useful and actionable. Police departments gain quantitative insights about traffic patterns on roadways, sidewalks, bike lanes, which they can apply to make better decisions about where and when to deploy officers or change roadway signage.
Police departments get more value from their existing video surveillance systems by complementing that investment with video intelligence software, which allows agencies to be alerted to situations in real-time, as well as uncover the data trends that are buried in their video surveillance data. This combination of technologies offers police agencies a variety of ways to enhance public safety for motorists, cyclists and pedestrians, while maximizing their staff resources and improving the accuracy and efficiency of investigations.
Signup to receive a monthly blog digest.