Abbas Mehrabian

Video analytics: What is it, and how can it deliver value?

Blog Post created by Abbas Mehrabian on Sep 26, 2018

Today, advances in technology allow us to capture digital videos through cameras as well as many camera-enabled devices, and petabytes of data are being gathered by these means every day. The challenge, then, is how to get the best possible actionable insight from all this data. This can be achieved through video content analysis (VCA) using a VA set of processing and analyzing tools.

Insights from video

VA enables companies to automatically monitor camera feeds, then send alerts or trigger actions based on an instant analysis. Using VA, enterprises can capture, store, transmit and analyze video data in real or near real time for business intelligence and actionable insights. The insights are in the form of patterns, behaviors, events and specific attributes so informed decisions can be made by the right people at the right time.

There are many use cases for video analytics based on specific business needs. Here are some of the more common ways VA is being used:

  • Security/surveillance. VA can be deployed to analyze feeds from cameras posted throughout a facility’s perimeter fencing to identify when an asset enters or exits an area. Notifications can be made when an asset has entered a restricted area, or an intrusion alert can be triggered if a person walks into a restricted zone. VA can also be used in a loss prevention capacity by providing comprehensive coverage of a retail store in areas of high risk, such as the point of sale and exits, with the goal of reducing inventory shrinkage
  • Smart cities. VA can be used to analyze traffic by automatically counting vehicles and classifying them as a car, bus, truck or bike. Other uses include analyzing parking spaces to detect occupancy, parking duration and vehicles parked in no-parking zones. VA can also be used to detect license plates on vehicles and match them with specific lists
  • Other uses. VA can be used for facial recognition (people identification, demographics), for detecting objects that have been added or removed from a location and for safety purposes such as monitoring a “person down” scenario. The technology has also been used for retail optimization by analyzing sales floor patterns and customer behavior. Other uses include scene analysis, behavioral and body analysis, crowd management and military functions such as force protection and guided attacks.

 

VA process and benefits

To get the best possible actionable insight from video data, it’s important to follow a typical VA process to capture and analyze the video images. The process includes tracking pixel movements and analyzing the overall image to identify “blobs.” In the analysis phase, pattern-based analytics algorithms or a graphic visualization can be created, and deep learning can be used for advanced data processing.

Technologies typically used for VA include sensors, cameras, software for data compression and network connectivity. Issues to contend with include privacy concerns and storing large amounts of video data. A key to VA success is effective data integration, and the typical process can be modified with other embedded processes to deliver optimal video images for VA.

Video analytics can deliver significant benefits, especially in the areas of security and public safety. The insights provided by VA can help enterprises better understand their physical environment and inform business decisions that provide value.

Outcomes