Long before we were surrounded by photographs via Facebook, Instagram and YouTube, we already knew a picture was worth a thousand words. If you’ve seen an image of a cheetah, you not only know something about its spots, but you also know it doesn’t look like a panther, or a tiger, or a Dalmatian.
That bounty of information we get from images and the comprehensive, instantaneous analysis we apply to them without even thinking, are skills increasingly mimicked by computers. Machine learning — the next step toward artificial intelligence — means computers learn as we do: gathering information in tiny bits, cross-referencing and comparing those bits with what else they already have in their “brain,” aka, database, and then generating a wealth of insights about it all.
These days, video from security cameras is one of the most common sources of information and one of the most underutilized sources of data. Most airports have an array of them already installed, as part of their increasing responsibility to ensure public safety and prevent crimes - or aid in their investigation after the fact. However, thanks to advancements in video analytics, these data-gathering and insight-generating devices can now be used to not only increase security, but to enhance travelers’ airport experience, leading them to have a preference for one airport over another.
Airports are no longer just a transit juncture we blast through on the way from here to there. They cater to travelers for longer intervals and build longer customer relationships, and are rising to the occasion of serving as enjoyable places to spend our time in between travels. Like shopping malls or city centers, modern concourses are filled with shops, restaurants, pubs and wine bars where travelers can enjoy a sampling of local culture, attractions and cuisines.
Because they are often travelers’ first experience of city, in many ways, shape people’s perceptions of the metropolises they’re located in, contributing to, or detracting from, their status and desirability as a destination. The insights gained from video analytics can help airports to capitalize on their new stature as a temporary home away from home, by improving safety, as well as the traveler experience by cutting wait times, and heightening enjoyment of shops and eateries.
As we at Hitachi developed our own video analytics suite, we learned a few things about what aspects existing solutions were missing that could improve accuracy, and deliver the insights and alerts our customers needed most. Here’s just a few of the ways airport administrators, security personnel and travelers can all benefit from video analytics:
A force multiplier for public safety
The top priority for any airport administrator will always be the enhanced security of travelers. Before they can appreciate moving through lines swiftly and enjoy some eating, drinking and shopping while they wait for a flight, they have to feel safe.
Fans of Sherlock Holmes or Jason Bourne can’t help but be awed by the numerous subtle observations that get stored in the hero’s vast mental database and then knitted together in an airtight version of events. Video analytics and intelligent IoT platforms work in the same way, taking data recorded by cameras and sensors, then sorting, distilling and analyzing it to predict the best course of action.
Good video analytics software can be optimized to detect the two primary hazards posed in airports: things and people. Object detection analytics can be used to provide airport administrators with alerts about inadvertently left-behind or suspicious objects, or intrusions into restricted areas. However, the simple pixel analysis offered by in early video analytics solutions cannot recognize objects in the field of view, only movement or disruptions. If the software only analyzes pixels—or if it draws its conclusions from the video version of a simple trip-wire, high rates of false positives will result. Security personnel will waste their time on unnecessary investigations and the alerts will be considered untrustworthy technology that cries “wolf!” too often.
In contrast, recent advancements in video analytics are now making it possible to reconstruct two-dimensional video images into 4-D understanding, by adding depth and time, which dramatically improves accuracy and reduces the number of false-positive incidents. With this capability, the software can better detect suspicious objects — even if they’ve been left outside the terminal in loading and unloading zones, poor visibility created by rain or snow won’t interfere with their identification. This is key because any deviation from a perfect scenario can confuse systems, reducing the accuracy and usefulness of the analytics.
In analyzing people to identify potential intruders, early video analytics solutions merely counted objects in motion, which meant they could also peg rolling luggage as a human. More sophisticated solutions are now calibrated to distinguish people whether they are sitting, walking or running, and to ignore interference such as shadows. Similarly, traditionally, these tools relied on pixels to detect color and frames per second, size and motion and broadly categorized forms as a person, car, animal or package, etc. By factoring in particles, perspective, velocity, path deviation and travel distance, today’s more advanced video analytics can now distinguish people from objects and animals, etc., with much greater accuracy. Day-to-day, alerts can be calibrated to account for actions like loitering, with a time limit that would recognize the difference between a person that is lost or pausing to make a phone call and one that is acting suspiciously.
Such sophisticated algorithms are useful to security personnel in two ways: Live face-matching allows for apprehension while a crime may be in process, and review of the historical footage can match that a suspect’s face to prior events, and increases the likelihood of identifying and apprehending them and their accomplices. The ability to quickly provide distributed situational awareness to an entire security team via mobile browsers further enhances the solution’s benefits and the team’s ability to keep people safe.
As the threat of attacks and natural disasters in public and private areas persists, security will remain a top priority. Video analytics provides a significant opportunity to ensure safety and serve as a force multiplier for law enforcement, first responders and airport security teams.
The end of the endless queue
In the past, travelers have considered airports less of a welcome oasis than a series of obstacles. Video analytics can help airport administrators turn that expectation around. Queue detection functionality can banish long lines, whether they’re occurring as curbside traffic jams in the drop-off and pick-up areas, or in security and passport control queues. Video analytics technologies can alert airport adminstrators when crowds begin forming, in real time. If queues pile up, they can deploy staff to open more ticket counters, security gates, or passport control lanes, which decreases wait times and increases customers’ satisfaction, freeing them to go on to enjoy an airport’s retail attractions. Viewed in aggregate over time, and by connecting flight information to the same system, this data can also help supervisors to predict demand, and optimize staffing plans and scheduling to ensure they can better support regular patterns of busy activity.
Smarter retailers make for happier customers
Online retailers are way ahead of brick and mortar stores when it comes to information about customers. They routinely get visitor behavior data from their websites and then rejigger the site content to determine optimal arrangement, imagery, and messaging that results in higher sales.
For airport retailers, video analytics now provides the ability to gain similar insights about the real world, so they can better reach their customers. Merchants can measure relevant details, like the ratio of people who enter to the people who walk by, or the level of interest in products and conversion rates at point of sale. They can also test advertising messaging and merchandise layouts for optimal performance: Which banners and mannequins attract the most interest? Where, when and for how long do people stop at different displays? If a store sells expensive items like designer watches, salesclerks can receive alerts when customers linger and then make sure those customers are waited on.
Beyond security and theft prevention, analysis of video footage can also help store owners adjust staffing for better floor coverage. Comparing it with register receipts and other data can helps them personalize customers’ experience of their shop. For airports looking to fill retail space, such data can attract retailers who want to run a smart business – the data can be used to identify high traffic areas to increase the rental value of the space.
How to get the most out of your video assets
As we have worked with our airport customers around the globe, we’ve consistently heard three main concerns: Will video analytics software interoperate with the systems they already own? If so, how will they protect their customers’ privacy? And how can they get the most out of their video security systems?
Each HVA analytic is engineered from the ground up, using a custom algorithm written specifically for each scenario, rather than relying on “rinse and repeat” generic modules or ones that cut corners and risk the possibility of false positives. The integrity of the physical camera system is ensured with the Camera Health Monitor, which automatically detects tampering or damage and will send an alert if issues arise to ensure valuable video data can be maintained as continuously as possible.
In addition to being accessible from a centralized operations center, the HVA and HVS system is distributed, so alerts can be sent to tablets and smartphones. TSA officers, for example, can receive the latest intelligence as snapshot images, GPS coordinates, and icons that deliver information about headcounts.
Beyond the terminal and runways
Video is the most underutilized source of data today, with the potential to provide a wealth of insights and real-time alerts that can help make airports safer and evermore appreciated destinations in their own right. Smart, data-driven airports allow patrons to enjoy quick movement through the system, and better opportunities to enjoy their time while they’re waiting for flights. Merchants can enjoy a captive audience with money to spend; and managers to enjoy a chance to outshine their competitors in nearby airport hubs.
Smarter airports have the potential to join smart cities as an important community member. One that is efficient and vibrant, where people enjoy spending their time while they travel. In coordination with smart cities airports and other organizations can share data and multiply their impact and improvements by coordinating together. By becoming smarter, airports set a great first or closing impression for travelers, and become a gateway to experience the smart city that awaits them.
Justin Bean is Director of Smart City Solutions Marketing at Hitachi Insight Group, which is responsible for driving the global unified internet of things (IoT) business and go-to-market (GTM) strategy for Hitachi, Ltd. Prior to Hitachi, Justin worked with startups and Fortune 500 companies that are applying IoT and disruptive technology to improve our lives and cities. He earned his MBA in sustainable management and lives in San Francisco, California.
