From Safe Campus to Smart Campus

Transforming campuses into safer and smarter institutions with artificial intelligence and data-driven solutions

Today, educational institutions operate much like a city, with centralized law enforcement ensuring the safety of students and visitors, responding and investigating when crime occurs and striving for improved quality of life for residents. Campuses house local businesses—sometimes even healthcare clinics and stadiums, as well—which can serve thousands of people daily.

Much like cities, campuses are embracing sophisticated artificial intelligence and datadriven solutions to understand, mitigate and solve security and operational challenges. Based on Machine Learning techniques, AIbacked technologies are trained to extract, identify, classify, recognize and index metadata from video surveillance content, which can then be leveraged as intelligence for driving anything from investigating crime to recognizing a security threat or emergency, and even deriving operational insights for improved business performance.

Video content analytics solutions, for instance, enable university and K-12 campuses to utilize their video surveillance infrastructure to extract data intelligence. Maximizing their existing investment in security surveillance, campuses leveraging video analytics are enabling increased public safety and accelerated security response, while also driving operational productivity and business performance for on-campus retailers, healthcare institutions, stadiums, mass event venues and more.

Video Content Analytics 101

Campuses are home to thousands of security cameras—these are installed on pedestrian pathways, main roads, around the perimeter of campus buildings and inside them and throughout parking lots. These sensors are used to surveil diverse spaces, from dormitories to classrooms, stores, stadiums, theaters, and both indoor and outdoor venues.

Security surveillance is designed to help campuses increase situational awareness, but—even with state-of-the-art cameras spread throughout the grounds—the challenges of reviewing and actively monitoring video are overwhelming. In the aftermath of an event, reviewing video across multiple cameras and recordings is both work intensive and inefficient for security. While the evidence extracted is valuable, dedicating the time and manpower to comprehensively comb through video is not always justifiable. Real-time manual monitoring of video feeds often is equally ineffective: Even the most attentive operators can be easily distracted or misinterpret a critical event.

To streamline the review of video and response to threats, video content analytics leverages machine learning to analyze video, extract objects that appear, identify and classify them and then index them as metadata. This metadata can be searched to pinpoint specific people or objects of interest or leveraged to configure alerts based on predefined rules. Investigators looking for a suspect wearing red can isolate all appearances of people wearing red in video, across all cameras and also trigger calls to action any time a person wearing red is detected in real-time in certain areas. Beyond this, by analyzing the aggregated data from long-term video recordings across cameras, law enforcement officers can uncover trends and define a benchmark for typical behaviors, so that unusual and suspicious activity is easier to identify.

Attaining Campus-wide Situational Awareness

The ability to configure real-time alerts can increase security’s awareness of on-campus activity and accelerate their response to threats. As soon as an alert is triggered, law enforcement can immediately assess the activity and determine whether something malicious is afoot. By bringing the most relevant details to the forefront, video content analytics alerts keep security focused and responsive.

While sometimes innocuous, oftentimes loitering can indicate a potential threat. For instance, school security might want to be notified of people or cars dwelling in a campus parking lot after hours, suggesting potential criminal activity. Video analytics can also help streamline daily school activities that require monitoring, such as student pick up and drop off—enabling security to take note of any excessive dwelling or unrecognized vehicles in the parking lot. During the school day, security can be alerted to outdoor movement during class time and quickly understand whether students are cutting class or whether interlopers are entering the campus. As active shooter situations continue to terrorize universities and day schools, the ability to immediately detect anomalous activity, attain situational awareness and respond to threats is critical.

Dwelling detection is also crucial for preventing on-campus theft. Campus safety experts note that loitering around dorm entryways, bike racks and facilities that store valuable equipment can indicate intent to commit a crime. With dwelling alerts for these areas, police can be quickly mobilized to deter or proactively respond to thievery.

Beyond dwelling, video analytics can be used to alert to overcrowding and traffic bottlenecks to mitigate the risk of violence erupting or the outbreak of a fight. On college campuses, where political gatherings can spark violence, police can closely track activity and react to indications of escalating risk.

Security surveillance is used across campuses, but with the addition of artificial intelligence and data-driven video analytics and real-time alerting capabilities, security can enhance situational awareness, detect unusual and excessive dwelling and accelerate responses to emergencies, threats and suspicious behavior.

Accelerating On-Campus Investigations and Security Response

On-campus crime is not always proactively preventable. In the aftermath of an incident, the primary concern is accelerating the post event investigation. Law enforcement must be able to quickly understand what occurred, identify suspects and then locate them to mitigate future threats and crimes.

Video content analytics enables campus law enforcement to accelerate investigations, enabling them to quickly and easily understand the event and extract evidence. When investigators have a suspect, they can search video based on known features of the individual, such as gender, clothing color, height—even the direction in which the suspect escaped.

In cases where suspect details are unavailable, investigators can leverage extracted metadata to understand quickly what occurred based on objects and people recorded in the video. Package theft, which is a common occurrence in dorm settings, can be easily investigated by isolating which people accepted package deliveries and by tracking suspects and packages’ movements throughout the dorm. Even when crime occurs over the weekend or at night, the operator can rapidly review hours of footage, filtering based on known criteria, to discover when the crime occurred, collect evidence and pinpoint suspects. This is equally effective when campuses must investigate accidents, vandalism or damages to school property, personal belongings or on-site vehicles.

Driving Operational Efficiency Across Campus

Over time, when video metadata is extracted, aggregated and visualized, it becomes a powerful, actionable asset for universities and K-12 campuses. Uncovering insight about how amenities are used and abused, operations managers can make data-driven decisions, develop contingency plans and streamline response to unexpected conditions.

Understanding movement in, out of and across campus, within and around buildings and along pedestrian and vehicular pathways, educational institutions can optimize campus- wide traffic and facility maintenance. The ability to evaluate which entrances and exits are most often used—and if any are underutilized— can help the campus proactively track maintenance and repair requirements and meet them before problems arise. Understanding how cars navigate within the campus, managers can insert infrastructure such as crosswalks or stop signs to increase pedestrian safety and prevent bottlenecks or accidents. Even within a building, the campus can address or prevent overcrowding by understanding usual trends and accommodating visitor preferences for navigating each building. A broader view of campus activities and entry and exit patterns also can drive traffic optimization, from coordinating public transportation, inter-campus transit and parking lots, based on need.

Beyond traffic concerns, understanding visitor movement and building occupancy patterns can drive more efficient energy consumption across campus. With video analytics, operations managers of sprawling campuses can quickly visualize where and when activity occurs and build sustainability strategies based on actionable data.

In addition to the visitors and students, campuses also oversee extensive staffs. Video analytics can help managers track employee activity, identify inefficiencies and develop solutions to drive productivity. Having granular intelligence about the goings-on about campus is key, and educational institutions should be taking advantage of all their resources to extract and leverage data for streamlining everyday functions and operations.

Providing Deeper Intelligence and Broader Value

Campuses are multifaceted and complex businesses, involving property management and rental concerns; event execution and organization issues; retail operations and performance indicators; among multiple other management functions. Leveraging security surveillance infrastructure, video analytics provides visibility and value across the broad organization, enabling the various stakeholders to derive granular insights for extensive business strategies. When it comes to renting space for on-campus retailers, restaurants and stadium vendors, the campus can extract data to demonstrate the strategic value of each property, to set fair and justifiable pricing. Beyond this, they can also support on-campus businesses with data intelligence generated from video surveillance. Stadiums and college sports teams can leverage actionable insights to optimize security, staff engagement, traffic and parking flows and concession placement and purchases. Retailers can prevent losses, abandoned purchases and disgruntled customers by embracing intelligence-based approaches to staff management, store configuration, security and checkout options.

The possibilities are endless and the common ground is video analytics. As campuses ready themselves for future security threats, they should consider technologies, such as video content analytics, which maximize existing infrastructure, provide broad value across the business and drive sophisticated and proactive security.

This article originally appeared in the March/April 2019 issue of Campus Security & Life Safety.

Digital Edition

  • Campus Security & Life Safety Magazine - March April 2019

    March/April 2019

    Featuring:

    • Proactive Steps in Protecting Students Using Technology
    • Securing Our Hospitals and Protecting Your Privacy
    • Leveraging a Unified Mass Notification Solution
    • From Safe Campus to Smart Campus

    View This Issue