Large event venues spend significant resources on physical security, and the vast majority invest in video surveillance systems as a way of monitoring public areas such as parking lots, garages and seating, as well as “backstage” areas such as loading docks and offices. Even though the camera footage is often preserved for investigating incidents, for the most part, video footage goes untapped because – realistically – security staff cannot constantly and efficiently monitor cameras in real-time or review the high volumes of archived video.
Video Analytics Powered by Artificial Intelligence
Increasingly, event venues are overcoming this challenge and maximizing this video data by investing in intelligent video surveillance. In this article, I’d like to explore how the combination of traditional surveillance and groundbreaking video content analytics enables event venues to not only optimize physical security and safety, but also improve customer service, streamline operations, and gather valuable business intelligence.
Deep learning and artificial intelligence-based video content analytic systems detect, identify, extract, and catalog all the objects that appear in video footage, based on classes and attributes such as gender, appearance similarity, color or size. This – along with behavioral detection, such as object interactions, dwell times, and navigation paths, as well as face and license plate recognition – enables searchable, actionable and quantifiable analysis of video. The technology makes it easy for users to accelerate post-event investigations, improve situational awareness, and derive critical business insights from video data.
Accelerate post-event investigations, improve situational awareness, and derive critical business insights
Accelerating Investigations and Real Time Response
Where there are crowds, it’s not uncommon for medical or criminal incidents to transpire. In the aftermath, security teams must quickly understand the scene and, by making video searchable, video analytics helps them efficiently gather information or evidence from multiple camera. Whereas a manual search of video footage would require hours, if not days, of valuable time and staff resources – while being subject to human error – a video content analytics system enables effective multi-video search using a variety of filters, so that post-incident investigations can be focused and, thereby, accelerated. For instance, if witnesses claim that a man in a yellow shirt caused a disturbance and then drove away in a blue pickup truck, security could filter video to review instances where people or vehicles matching those descriptions appeared, and then focus their investigation on relevant video evidence.
For the same example, if the perpetrator was still at large in the venue, security managers would want to track the person’s movements and, ultimately, apprehend him. Using the same cataloged metadata that drives video search, intelligent video surveillance accelerates real-time response to developing situations from threats and suspicious behaviors to emergencies, or even any behavior that requires closer monitoring or intervention. Something as mundane as an unwieldy queue forming at a concession stand, can translate into alerting logic that increases operators’ situational awareness and ability to effectively intervene when time is of the essence. By understanding developing situations in real-time, they can respond more quickly to prevent or solve a problem.
Rule-based real time alerts
Users can configure rule-based real-time alerts, based on specific scenarios – such as when people counts in the defined space next to the cashier are higher than the specified threshold, indicating a queue, or when vehicles are detected dwelling in a no-park zone or sensitive security area. In the case of the offender in the yellow shirt, security could configure alerts to notify for appearances of men in yellow shirts or blue pickup trucks within the event complex. If they’re able to extract the perpetrator’s face or license plate from the video search, security officers can also identify and track the specific person and vehicle of interest.
Intelligent video surveillance users can compile “watchlists”
Object recognition is helpful for tracking specific, identified threats, but it can also be leveraged for locating missing persons or vehicles or offering VIP services to guests who comply. Intelligent video surveillance users can compile “watchlists” – that include persons of interest, such as past agitators, recognized criminals or terrorists, missing persons or VIPs – to be alerted when matches for these faces appear in video. Face recognition can also be used to differentiate recognized employees from visitors that are not authorized to be in certain spaces, to prevent security breaches.
Because camera conditions aren’t always optimal for “in the wild” face and license plate recognition – where lighting conditions or camera angles don’t always allow for a pristine face or plate capture – it is critical to have a comprehensive video content analysis tool for triggering alerts based object classifications and non-personally-identifiable parameters is critical.
Optimizing Operations Based on Business Intelligence
While having real-time situational awareness is crucial for proactive intervention to developing situations, understanding long-term activity and trends can be equally important for an events venue or stadium. With video analytics, video surveillance data can be aggregated over time to generate statistical intelligence, trend reports, and data visualizations for operational intelligence. This capability transforms video surveillance from a siloed security tool to an intelligence platform with broad value for business groups across the organization: Marketing departments can gain visitor demographic information for targeted advertising and product placements. On-site retailers can uncover how visitors navigate their stores and the products or spaces that draw the most attention. Property management can follow heatmaps demonstrating where visitors dwell and the durations they stay at certain exhibits or areas, to base rental and leasing fees on actionable and quantifiable intelligence. Operations managers can more effectively plan and hire staff for events, by analyzing visitor traffic trends, entry and exit points and occupancy statistics.
Video surveillance data can be aggregated over time to generate statistical intelligence
For these reasons and more, event organizations that already rely on CCTV networks, should consider how to better leverage those existing investments to deliver substantial value across the organization: By making video searchable, actionable and quantifiable, intelligent video surveillance empowers event managers to enhance public safety and physical security, while also streamlining operations and improving customer service.