AI AND VIDEO ANALYTICS BLOG
Video Surveillance & Physical Security Industry Viewpoints
December 30th, 2019
Author: Marc Sanders

Increasing Physical Security with Video Analytics for Entertainment Businesses

Intelligent Analytics for Theme Parks

Physical security is a priority for theme parks, especially since so many cater to families with children, with major theme parks drawing hundreds of thousands, if not millions, of visitors each year. Such entertainment complexes require 24/7 security to safeguard the attractions as well as the visitors and staff.

Theme parks contain multiple complex public areas, such as parking lots and food courts, that need to be patrolled by staff or monitored remotely using cameras. But there are also employee-only, or behind-the-scene areas that must be monitored; such as administrative offices, areas where park mechanics are operating (think roller coaster rides) and sensitive exhibits (exotic animals and wildlife living and care facilities, for instance).

To support security operations, entertainment businesses typically deploy video surveillance cameras; however, in reality, the footage captured by the cameras is often dismissed as routine requirement: It exists mainly to support post-incident investigations for the security department or local law enforcement. This is because manual review of video footage can be tedious, time-consuming and prone to human error. Fortunately, in recent years video content analysis software has made it possible to quickly, easily and effectively analyze and review such footage, as well as to receive real-time alerts for detected anomalous behavior and to discover actionable business intelligence that is highly useful for operations, security and retail managers.

Incident Analysis and Investigation

Theme parks are concerned with preventing theft, loss, damage, or injury throughout their extensive estates, which include retail gift shops, food vendors, exhibits, live events and more spaces that must be secured. When a security breach or incident occurs in one of these domains, security staff often start investigations by reviewing video footage from the scene. Powered by Deep Learning and artificial intelligence, a comprehensive video content analytics solution can process video, identify objects in the footage (including people and vehicles), and index these objects for effective rapid search and analysis pinpointing specific objects of interest. Hours or days of footage can be combed in a matter of minutes, with operators filtering video to focus the video search and show only people or objects that fit a certain description. This empowers security teams to extract crucial details, gather evidence and accelerate investigations.

Real-Time Alerts for Situational Awareness

Video intelligence software can be used to deliver alerts in real-time, so operations or security managers can act swiftly when anomalous behavior or events are detected. A human operator defines the rules that reflect what is “normal” and configures alerts to be triggered when objects or activities outside the scope of “normal” are detected. Examples of anomalies could include anything from unusually long queues at a concession stand to a person trespassing in a defined sensitive area which is off-limits to unauthorized personnel. For instance, count-based alerts can be triggered when the number of people detected crosses a certain threshold; anything in excess of a certain number will send an alert to the correct authority, who can then determine whether a reaction is needed and dispatch security or personnel resources to respond as required.

Facial Recognition to Prevent Unauthorized Entry and Enable Search

To prevent unauthorized persons from entering a park or restricted sections of the park, security managers have the ability to upload watchlists of video or digital images, enabling video content analytics software to automatically monitor all camera feeds for face matches. When a possible face match has been detected, human operators receive a visual/audio alert, and can validate a correct match. Based on the human evaluation, the detected person can be more closely monitored and tracked, when necessary, confronted to prevent him/her from entering the park grounds. Real-time or recorded camera footage can also be searched post-incident, based on specific faces of interest to advance an investigation and extract video evidence.

Analyzing Operational Trend Statistics

To ensure the safety of venue visitors and staff, security teams need to tap into quantifiable data about visitor traffic flows available within their existing CCTV platforms. Video content analytics collect and aggregate data over time, and visually present this in reports, dashboards and heatmaps, revealing trends that both security and operations managers can use to understand pedestrian and vehicle patterns, and proactively anticipate situations. For example, the data can visualize the number of visitors detected at an entrance or exit during different times of day – so operators can understand the most popular times for arriving and leaving the park – coupled with the ability to extract park occupancy statistics for specific days, weeks, or months. Heatmaps can indicate where visitors commonly walk, which exhibits they view for the longest durations, which entrances/exits/restrooms see the most foot fall, and how many cars and pedestrians enter specific parking facilities.  This critical business intelligence can help theme parks understand whether some entrances and exits aren’t fully utilized, or identify where crowding regularly occurs and determine strategies for reducing bottlenecks. Empowered by this information, security teams can prepare for expected traffic peaks by hiring more staff, or proactively redirecting staff to the high traffic areas when an increase in people is detected.

By taking all this data into account, the leisure and entertainment business can actively improve public safety and the visitor experience through clever investment and application of CCTV analytics, thereby increasing the potential of return custom.

Most entertainment venues rely heavily on video camera surveillance, but to fully leverage that investment, there is a very real operational requirement to effectively analyze video footage and turn it into actionable intelligence. Video content analysis is a logical extension of video surveillance that dramatically improves efficiency and accuracy in security operations. But – as a future blog post will discuss – video surveillance can be maximized for so much more when entertainment venues derive business intelligence from their video system to empower other departments such as marketing, retail and operations.