As the world grapples with the COVID-19 pandemic, it has become clear to public health officials that physical distancing, wearing masks, and contact tracing are essential to prevent the spread of the disease. Encouraging, monitoring, and enforcing those public health directives is a daunting task, especially for public transit agencies serving thousands or millions of people each day with bus, rail and subway lines: In New York City, for example, 4.3 million people ride the subway daily. How can transit agencies prevent crowding, monitor mask-wearing, conduct contact tracing among their employees, and optimize the cleanliness of their facilities and equipment as they resume service?
To solve these challenges, transit agencies can leverage their existing video surveillance networks by pairing that technology with video intelligence software. Video content analysis that is powered by Deep Learning and Artificial Intelligence extracts, identifies, classifies and indexes valuable video metadata and presents it in ways that make it searchable, actionable, and quantifiable. With video analytics, transit managers and their security teams can obtain trend data, as well as receive real-time alerts to respond to evolving situations such as crowding.
Crowding leaves no room for social/physical distancing. While transit agencies and facilities are tasked with preventing crowds, the very nature of rush hour is a crush of people crowding onto tight hallways, platforms, trains and buses. Turnstiles slow the flow of traffic into a transit station, but on their own they can’t indicate whether a crowd has formed or a station has exceeded capacity. To ensure that facilities and public areas are not crowded, managers must be aware when the number of people in an area exceeds a specific threshold.
Video content analysis systems can be configured to detect pre-defined anomalies based on traffic patterns, norms and benchmarks, and the technology can count people in a station, parking lot, or other space to increase occupancy awareness. When the system operators set a custom threshold, they can trigger people-counting alerts to raise awareness when pedestrian traffic has exceeded the norm and determine how to reduce occupancy to increase public safety. Security or operations team members can deploy additional personnel or signage as needed to redirect traffic to other walkways or access points to alleviate crowding. Video analytics data can also be aggregated over time and visualized in the form of customized reports and dashboards, including graphs, heatmaps, and histograms for demonstrating patterns and trends, so that operations managers can benchmark norms, make intelligent decisions and develop contingency plans.
In the COVID-19 crisis, it has become critical to detect, monitor and analyze distance between individuals, because the virus is so contagious. Video content analytics technology can detect the proximity between people, enabling video operators to search and filter video to see where and how often physical distancing was violated. Transit agencies can apply this capability to observe whether their constituents are observing physical distance mandates, and identify when and where enforcement is needed. Because the technology can discern when two or more people are standing within a pre-defined proximity, custom rules can be set up to trigger alerts for distancing violations, so security can respond proactively and preventatively.
Just as importantly, operations managers need to be able to demonstrate compliance with physical distancing mandates and identify problem hotspots. By aggregating long-term data and creating dashboards that show peaks and valleys of traffic over hours, days or weeks, operators can obtain a holistic view of information, such as the average distance between pedestrians, times of day or week when physical distancing was more (or less) common. Operations staff can then decide whether operational or staffing changes are needed to support physical distancing, such as changing transit schedules, or entrances and exits.
If a public transit employee self-identifies to their employer that he or she has tested positive for COVID-19, video analytics can be used to conduct a forensic video review for identifying the employee’s interactions with others while he or she was possibly contagious. The employer can upload a photo of the employee to the video analytics system and use facial recognition and multi-camera search to pinpoint the employee across locations and whether any proximity violations with others took place. By generating a list of people who encountered the infected employee, the employer can protect the person’s anonymity, while informing those who interacted with him or her that they need to self-quarantine. They can also assure those who did not have contact with the infected person, so they can be more at ease.
In cases where face recognition cannot be used, video search can be conducted across cameras based on the description of the employee’s clothing, using an appearance similarity filter. This can help the employer identify the areas he or she visited and at which exact time, to understand which other employees may be at risk.
Another key filter for contact tracing is face mask detection – the ability to productively identify a person’s interactions and whether either or all the people involved were wearing masks, is important for assessing risk of contagion. More and more businesses, workplaces and state agencies are mandating that people wear masks whenever physical distancing is not possible. Employees in a public transit facility may be required to wear a face mask to protect the public and their co-workers. To enforce that workplace safety mandate, supervisors need the ability to detect employees who are in violation of that rule. Video content analytics systems can detect whether a person is wearing a mask and can be configured to send a real-time alert whenever it detects someone who is not wearing a face mask to allow for swift intervention.
Transit operators can also gather data over time to analyze the face mask adoption rate (for both riders and employees). In terms of employee compliance, this is an important way to provide evidence of employer compliance to government officials. Transit managers can also analyze mask trends in relation to time or location, and to identify common hotspots where visitors or employees are not wearing face masks. These metrics may also help transit managers and government authorities determine where to deploy staff or signage to remind people to wear face masks, and whether enough people are wearing masks to make it reasonably safe for riders to be on a subway, bus, or train.
Cleanliness and sanitation are keys to public health, especially when it comes to subway, train and bus stations and vehicles, which experience constant and heavy rider traffic. With count-based alerts, video analytics can help operations teams align sanitation and maintenance efforts with actual facilities usage instead of traditional time-based cleaning scheduling. By triggering customized alerts based on people counts and line crossing rules for area entryways, hallways, or other highly trafficked areas, public transit maintenance managers can more effectively and proactively clean spaces, as needed, based on actual traffic.
As with other video analytics, in addition to real-time alerts to facilitate immediate response, operators can build dashboard reports to understand traffic trends and maintenance requirements over time and across multiple locations. Heatmaps can demonstrate where people tend to walk and graphs can illustrate how many times people go into specific areas such as a restroom or food court. Operations managers can then predictively hire cleaning and maintenance staff based on movement trends and patterns.
Beyond maintenance and usage statistics, occupancy restrictions are another compliancy that transit managers must consider to keep visitors and employees safe and healthy. Using people counting and occupancy analytics, building managers can know how many people are currently on the premises and trigger alerts when customized thresholds are met.
For certain times of day or popular travel holiday times, transit managers can be better prepared for crowding if they have quantifiable, historical data regarding where and when people typically use transit stations. Video analytics provides such occupancy detail over time and location so that transit operators can see the trends and plan to allocate security or customer service resources.
Mass transit is essential to the economy, and good for the environment; people depend on it to get to and from their work, and it reduces car traffic. During this public health crisis, transit agencies are under tremendous pressure to make their facilities and equipment safer and cleaner for their passengers; to do so, they must optimize their sanitation operations, conduct contact tracing among their employees, and monitor physical distancing and mask wearing. Transit agencies typically have thousands of video cameras in place; video intelligence software is an effective way for them to make mass transit cleaner, less crowded, and safer.
Signup to receive a monthly blog digest.