Automating Anomaly Detection in Airports, Subways, and Public Places

Airports, subways, and public places fall victim to various malicious events like terrorism, vandalism, graffiti, gang activity, and more. Most of these areas are fixed with extensive video surveillance systems. However, the detection of any suspicious activity is wholly dependent on the humans who are watching the monitors for any signs of abnormal behavior. 

There are several problems with this approach. For starters, humans are fallible. They can make mistakes and miss important and dangerous events. Such slip-ups can lead to severe consequences that can adversely affect such public places. For instance, think of a small fire that is about to break out in an airport. If the video footage is being monitored by security officials, there is a high chance that they will not see the smoke and fail to report it. This could potentially set the whole area on fire. 

In such a scenario, we can say that our public security is completely dependent on the officials who monitor safety footage. But how efficient is this? 

By using video analytics, authorities can prevent such incidents from happening to a large extent. These cameras make use of software algorithms to identify pre-defined activities and scan for suspicious individuals. For example, cameras with facial recognition technology can help authorities find terrorist threats and detect individuals on government watch lists. Through behavioral recognition technology, cameras can be programmed to detect incidents such as left objects, congestion, reverse movement through checkpoints, and cars spending too much time parked in one spot outside of the building. All of these capabilities can help increase the safety of our public spaces, airports, and subways.

In this article, we will be discussing the possibilities of anomaly detection for airports, subways, and public places. Before we begin, let me walk you through the below topics.

• What are anomaly detection systems?
• How do outlier systems work?
• Anomaly detection with machine learning
• Where can we use such systems?

What are anomaly detection systems?

Although most public areas are equipped with surveillance throughout, most traditional systems are unable to detect and report security threats in real-time. This can cause some serious safety concerns in such areas.

Anomaly detection systems can detect any irregularity or peculiar behavior in public places or in any data set to detect any threats or future perils. As soon as a threat is detected, they immediately send a notification to the concerned authorities, so that the required action can be taken. Many companies have started using these systems for elevating protection in public places like subways, shopping malls, airports, sports stadiums, and more.

Also known as outlier analysis, these systems can be a lifeline for any organization or area’s security. They can identify abnormal or suspicious behavior and immediately channelize a scrutinous approach towards the beings and objects.

How do outlier systems work?

Generally, public places like subways, airports, shopping malls, etc. function under CCTV surveillance to ensure security for the public. Authorities install cameras, fire sensors, body heat sensors to detect any anomaly. Behavioral screening plays a major role in these discernment systems.

The footage from these cameras is being monitored from a control room and security personnel must stay alert at all times to detect any movements. Unfortunately, in these cases, human error is unavoidable as human beings can be easily distracted or tired from watching a screen continuously.

Complete automation and artificial intelligence can overcome this human limitation. Together, the human brain and computed intelligence can ensure a complete blanket of security and safety.

Recent progress: Anomaly detection with machine learning

Solutions with an anomaly detection feature can easily identify suspicious behavior, which is extremely beneficial in threat handling and assistance support. Once the system detects a threat, it can send a notification to the concerned authorities so that the required corrective action can be taken. By using a solution like this in public areas, authorities will be able to curb public incitement. Mobs can be easily managed and public nuisance can be handled easily.

Machine learning algorithms can be trained to detect any deviation from the normal path with training data. The primary difference between conditional video streaming and artificial intelligence surveillance is the level of security and safety. A machine learning incorporated area can manage the increased mass of crowds and streamline crowd management. The classical detection methods include distance base, neighbor based, ensemble-based, domain-based, and reconstruction based security anomaly detectors. Nowadays, it has been integrated with machine learning, advanced alarming systems, and other new technologies.

Where can we use such systems?

• Tackling terrorist activities – Public areas are a hub for terrorist activities. Places like subway stations, hotels, malls, etc. are high-risk targets for terrorists. By installing an anomaly detection system in such areas, it will be easy for law enforcement to spot and thwart terrorist plots.
• Crime investigation – In most criminal investigations, video footage acquired from CCTV cameras plays a major role in gaining evidence. Video surveillance footage can help authorities investigate incidents and identify suspects as well. Video surveillance footage can prove extremely useful in investigating incidents where a crime or terrorist act is carried out.
• Prevents vandalism– Anomaly detection systems can help you identify acts of vandalism like graffiti, spray painting, property destruction, etc. This can keep your premises completely free from any property destruction.
• Smart camera functionality – Intelligent IP cameras often feature built-in video analytics such as behavioral and facial recognition. Installed in a subway or train station, such cameras can help to identify potential threats and spot suspicious activity such as objects left unattended.

Ever since the airports have come into action, airports’ anomalies have been a serious security threat.

Be it the Glasgow airport attack in 2007 or the Karachi airport attack in 2014, then seven years of gap gave us no alarm for security reasons. Both incidents happened when terrorists or attackers successfully escaped the airports and detectors’ security alarms, the security system poorly failed.

Such scenarios can happen today as well but can be curbed if anomaly detectors are used well. The system should be well integrated with artificial intelligence to make sure nothing sets the aviation industry aback.

A sustainable system should be developed, and checkpoints should be installed.

However, both the attacks were purely sudden and demanded immediate implementation of backup plans, giving us another alarm for being ready with backup plans as well that too the ones with higher efficiency and amendments.

A careful anomaly detection system should be installed that can detect explosives such as guns, granites, knives, guns, etc., and can further alarm the system to channelize security measures at a time.

To conclude

Anomaly detection serving security and identifying data breaches in public places are open for research for statistics integration, process control, system integrations, software development, and many more.

Also known as an outlier, an anomaly detection system, is prone to hacking and is therefore often integrated with cybersecurity cells to establish breach-proof software or platform that can detect, alarm, and can plan security events.

The crowded places such as subways, malls, stadiums, airports having a large number of people along at a time should be equipped with all security gear and backup plans to save masses from any mishappenings or poor conduct.