31 Oct Enhancing perimeter security with AI
Perimeter protection is a crucial aspect of security, whether it be for public, residential or commercial purposes. Ara Ghazaryan, Ph.D, Technical Co-founder and VP of AI at Scylla, looks at how it can work even harder thanks to artificial intelligence (AI).
Even in safe regions such as the Gulf, security is a factor that really matters. It ensures that organisations, their assets and information are well protected from theft or damage, that nothing will disrupt the continuity of their operations, and that both employees and customers have a safe and comfortable environment.
Perimeter security is the first line in multiple layers of defence, protecting the boundaries of organisations, car parks, entry areas and loading zones. For many years, physical barriers such as fences, walls, doors and guards have been used to patrol the perimeter and monitor any suspicious activity. For example, the Al Fahidi Fort in Dubai has protective walls and watchtowers to provide visibility of the outside environment. Similarly, in Saudi Arabia, the Masmak Fort, with its thick walls, watchtowers and narrow, heavily built gates, protected against unauthorised entry and provided facilities to support ongoing life inside.
However, physical barriers have their limitations and are not sufficient to meet the growing demands of modern security threats. These methods are expensive to maintain, require significant manpower, and are not foolproof. Intruders could still find ways to breach the perimeter undetected. Besides, they do not provide much information about who is trying to breach the perimeter.
Advancing technology
With the advancement of technology, there are now various solutions available to enhance perimeter protection, including video surveillance cameras linked to a Security Operations Centre and access control systems that use security cards, PIN pads, biometrics, and other sensors. Infrared cameras are capable of capturing images in low-light conditions, making it easier to identify intruders. Laser beams create an invisible barrier that triggers an alarm when breached. Radar systems detect objects and movements beyond the line of sight, enabling early warning and detection of potential threats. Motion sensors detect movement within a specific area and send an alert to security personnel.
The major issue with video surveillance is its reliance on human review. This is labour-consuming, costly and fatiguing for security officers, which often results in missed events. A commonly cited study from 2002 found that after 12 minutes of continuous video monitoring, an operator will often miss up to 45% of screen activity. After 22 minutes of viewing, up to 95% is overlooked. One solution is to set up perimeter lines and detect line crossings or show the camera feed when there are a certain number of changed pixels.
Unfortunately, this approach doesn’t work well, and many detections end up being false alarms.
Avoiding false positives
Advances in AI technology have improved the efficiency and effectiveness of intrusion detection and perimeter protection solutions. This means that they can provide more accurate alerts. Due to false alarm filtering technology, it’s now possible to remove over 99% of false positives caused by insects, camera issues, lighting changes or small animals. Only footage showing a person or vehicle is displayed. This saves countless hours for security officers, reduces labour costs and minimises security operator fatigue.
By leveraging AI, organisations can enhance perimeter protection capabilities. AI-powered solutions use machine learning algorithms to analyse video feeds from surveillance cameras in real-time. They can detect and alert security personnel to potential threats, including unauthorised intrusions, suspicious behaviour, and unusual movements, such as individuals loitering in a certain area or objects left unattended. By applying machine learning techniques, these systems can identify patterns and anomalies in the data that may indicate a potential intrusion.
This helps minimise the risk of a breach and allows security personnel to respond quickly and effectively.
The power of integration
One more important step in improving perimeter protection is integrating access control systems with facial recognition software and implementing multi-factor identification to verify both employees and visitors. It results in increased safety and security and improves employee attendance control and visitor management. Using license plate recognition systems is another way to provide an extra layer of security. Along with video surveillance cameras, organisations can control access to their premises by creating a list of vehicles allowed on the territory. The LPR system quickly and accurately verifies vehicles that enter or exit a parking lot or a controlled area within a business or private premises, ensuring that only authorised vehicles have access to these areas.
Another prospective technology for perimeter protection used in the Middle East is security drones equipped with high-resolution cameras. They allow for quick and efficient monitoring of large areas, providing real-time footage to security personnel. This technology is particularly useful for monitoring areas that are difficult to access on foot. For example, oil & gas companies can effectively inspect pipelines that span thousands of kilometres or monitor the exterior surfaces of storage facilities. In combination with advanced AI solutions, surveillance drones are going to become smarter in the future and provide more high-quality data. Thus, in the next three to five years, the oil & gas industry is expected to increase its investment in drones and robotics from 15% to 28%, according to Frost & Sullivan.
Adoption across industries
The adoption of AI-powered video analytics for perimeter protection is already underway in various industries in the Middle East region. It is particularly useful for large-scale facilities, such as airports, seaports, oil & gas installations, and critical infrastructure, where human monitoring is impractical or insufficient. The Dubai Police Force has deployed a smart surveillance system that uses facial recognition technology to identify wanted criminals and suspects in real time. The system has helped the police reduce crime rates and improve public safety in the city. Similarly, the Abu Dhabi National Oil Company (ADNOC) has implemented an AI-powered video analytics system to monitor its oil and gas pipelines and prevent theft and sabotage. Airports use it to secure their perimeters and prevent unauthorised access to restricted areas. Banks protect their branches and ATMs from robbery and vandalism using technology.
Strong perimeter security is of utmost importance for manufacturing plants, which are at risk of product tampering and equipment theft, as well as vandalism and setting fires. Energy sites have to be vigilant to prevent terrorist attacks and acts of sabotage. AI-powered solutions help them monitor the area 24/7, detecting and tracking potential threats and automatically sending alerts before intruders enter the perimeter or do damage.
With the travel and tourism sector welcoming 90 million international arrivals per year, tourism is continuing to grow in the Middle East. Visitors are attracted to a warm climate, a safe tourist environment and sightseeing opportunities. Therefore, Middle Eastern countries will gradually start requiring the implementation of AI video analytics in malls, residential areas, recreational facilities, and hotels to prevent intrusions and increase security protection. Perimeter security is also important for unattended swimming pools, as there have been many incidents of drowning when people have entered a closed pool without lifeguards.
The right combination
No single technology can provide complete perimeter protection. Therefore, a combination of traditional measures and innovative technologies should be used to ensure maximum security. Although traditional measures are still prevalent in many parts of the MENA region, innovative technologies are becoming increasingly popular due to their effectiveness and efficiency. The adoption of AI-powered solutions is gaining momentum and it is likely to be only a matter of time before it is used widely.