Video analytics and monitoring tools are one of the best ways to keep cities and businesses safe. Public spaces such as parks and city streets can use video surveillance to track and record unusual activity.
With video surveillance, it’s easier to spot crimes, traffic congestion or accidents, and other issues that need to be addressed. But video monitoring also takes a tremendous amount of human and technology capital. To be effective, video systems need:
Because of how many video feeds are needed, human monitors often end up juggling multiple video streams, making it difficult for them to notice important situations – like an accident or burglary. The Nokia IoT Scene Analytics solution solves these problems by leveraging machine learning, pattern recognition, and behavioral knowledge to:
Download The Scene Analytics Application Note
The Nokia IoT Scene Analytics technology takes video analytics to the next level. Traditional video analytics systems, for instance, don’t work well in real-life environments. They rely on less adaptive, rules-based analytics and objection recognition software that can’t understand complex environments. That puts intricate traffic patterns, pedestrian movements, and city layouts outside of their scope, often causing false positives.
Because they don’t work well in real-life environments, it’s not practical for municipalities and businesses to apply traditional video recognition technology in highly variable and complex situations, like a live city environment with train stations, bus shelters, airports, and other public areas.
Real-life situations require an intelligent platform that can proactively detect, select, and track only relevant video streams for a variety of surveillance tasks.
The Nokia IoT Scene Analytics solution works by turning cameras into IoT sensors. It derives abstract information such as motion, direction, density, and velocity vectors from each video feed, then uses machine learning technology – developed by Nokia’s Bell Labs – to analyze these vectors in real time. The software establishes patterns, identifies anomalies, and generates alerts.
Nokia IoT Scene Analytics enhances situational awareness and prioritizes streams for their anticipated relevance, automatically allocating network resources based on dynamic application needs. The least relevant streams are cut off at the source, ensuring network availability for the most critical ones.
The Nokia IoT Scene Analytics solution, in parallel with the Nokia IMPACT IoT Platform, works with any video monitoring solution. The software is agnostic to video resolutions, camera brands, and networking options, so you can quickly implement it with the equipment you already have.
Scene Analytics is valuable anywhere you need to monitor complex, active scenarios in everyday environments.
For example, a camera monitoring a highway could be programmed only to transmit a video feed when unusual activities occur, such as a traffic jam, the presence of a pedestrian, an accident, or when a vehicle is traveling in the wrong direction. By transmitting video only when a significant incident occurs, human monitors don’t waste hours watching traffic flow smoothly.
Many cities also use video analytics with monitoring systems from public areas, such as a street lined with shops. Instead of having human monitors watch the street 24/7 and storing hours of video feed, they could use video analytics to detect anomalies. For instance, Nokia IoT Scene Analytics would automatically detect a store break-in and highlight that video for monitoring personnel.
For more information on when to use Scene Analytics and its features, download the free application note.
"The Nokia IoT Scene Analytics solution leverages machine learning, pattern recognition, and behavioral knowledge to anticipate what’s relevant, based on the prescribed video surveillance task."