CCTV Monitoring

No. CHeKT removes the need for port forwarding and DDNS. It secures remote video monitoring by eliminating port-forward rules and DDNS names, giving you a safer alternative without requiring static IP addresses.

CHeKT uses Cloud Relay to create a secure, encrypted TLS connection between CHeKT Video Bridges and the CHeKT Cloud. This connection encrypts video and securely authenticates users.

The CHeKT Cloud makes it easy to manage access and password security from one central platform. CHeKT encrypts all video streams and includes a patented privacy mode that limits monitoring agents’ access while they deliver services. CHeKT also runs on Amazon Web Services and complies with GDPR, further strengthening overall security.

CHeKT does not use a traditional VPN. Instead, CHeKT uses the TLS protocol to establish secure communication channels through its Video Control Panels. This approach gives users encrypted access to video devices without requiring VPN infrastructure.

Our interactive video monitoring platform is an open and scalable solution designed to seamlessly unify your security infrastructure. It offers robust integration with over 170 leading manufacturers, supporting third-party VMS platforms, NVRs, DVRs, encoders, and cameras.

By connecting your video surveillance system with our centralised monitoring platform, we can provide a comprehensive suite of services:

* Video Verification: Live, pre-, and post-alarm imaging for rapid and accurate threat assessment.

* Video Patrols: Flexible ad-hoc or scheduled remote visual site inspections.

* Device Uptime Monitoring: Immediate alarm notifications anytime a device loses connectivity.

* Camera Health Diagnostics: Camera failures, hard disk errors, and video loss are automatically handled and escalated as priority alarm signals.

* Remote Access Management: Secure, remote access control for employees and authorized personnel to customer facilities.

* Door Station Answering: Remote intercom communication and entry management.

* Automated Triggers: Seamless activation of relays and system presets based on specific events or schedules.

AI Video Monitoring can significantly reduce false alarms, but cannot eliminate them.

How AI Video Monitoring Reduces False Alarms:

AI-Based Filtering: AI Video Monitoring uses AI video analytics to filter out “False Positive” alarms by distinguishing between genuine threats and benign triggers like moving trees, animals, or changing light conditions.

Advanced Object Classification: AI video analytics can classify detected objects (humans, vehicles, animals, etc.) to determine if an object is a threat.

Cloud-Based Continuous Learning: AI algorithms continuously update and refine their detection accuracy, helping to reduce false triggers over time.

Scene Noise Cancelling: AI algorithms filter out environmental noise, such as reflections, weather changes, and small moving objects, reducing unnecessary alarms.

Limitations

No AI is Infallible: AI systems can still misclassify threats or fail to detect actual incidents, similar to how AI, like ChatGPT, sometimes provides incorrect information.

Environmental Challenges: Poor lighting, camera placement, and occlusions can affect detection accuracy.

Potential Missed True Positives: AI can sometimes fail to detect security threats due to biases in training data or real-world complexities.

Conclusion

While AI video monitoring cannot eliminate false alarms, it can drastically reduce them, enhance security operational efficiency, improve response times, and minimise operator fatigue. Human oversight and complementary security measures remain essential for optimal performance.

Using video analytics, including object detection, classification, and behaviour analysis, AI differentiates between human movement and environmental factors like rain, wind, or animals. Here’s how it works:

Object Detection & Classification: AI-powered surveillance systems use deep learning models to identify and classify objects in video. These models, trained on extensive datasets, can distinguish between humans, vehicles, animals, and non-relevant environmental factors like rain, shadows, or moving leaves.

Behavioural Analysis & Motion Filtering: AI analyses movement patterns to determine whether an object’s motion is characteristic of a human (e.g., walking upright, consistent stride patterns) or something else (e.g., erratic movement due to wind, birds, or reflections). For instance, Evolon’s AI algorithms can differentiate between people and vehicles while filtering out motion from rain, shadows, and small animals.

Scene Noise-Canceling & Environmental Adaptation: AI systems apply scene noise-cancelling techniques to reduce false alarms caused by environmental conditions such as fog, wind-blown objects, and shifting lighting conditions. These algorithms adapt to changing environments, ensuring reliable detection even in outdoor settings.

Machine Learning for Continuous Improvement: AI-enhanced surveillance solutions continuously learn from real-world data. This allows the system to improve accuracy over time by refining its ability to differentiate between true security threats and false positives.

Combining these capabilities, AI-based video analytics can significantly reduce false alarms while ensuring that legitimate security threats, such as unauthorised human movement, are accurately detected and escalated.

A false positive in AI video analytics occurs when the AI incorrectly identifies an event, object, or action as a security threat when no actual threat exists. The system generates an alarm based on this incorrect interpretation of video data.

For example, a false positive might occur when:

The AI mistakenly classifies a shadow, moving vegetation, or an animal as a person or vehicle.

Rain, snow, or reflections trigger an alarm due to misinterpretation by the system.

An object, such as a sign or a tree, is incorrectly identified as a human or vehicle.

False positives are a critical issue in video analytics as they can lead to “false alarm fatigue,” where staff may start ignoring alarms, potentially missing genuine threats.

AI handles occlusions—such as an object blocking the camera’s view—by using various techniques within video analytics and artificial intelligence. Here’s how AI-based video monitoring systems mitigate occlusion challenges:

Object Persistence and Tracking: AI-based video analytics use object tracking algorithms to predict and maintain the location of partially or temporarily occluded objects. By analysing an object’s movement before and after occlusion, AI can infer its probable position even when not fully visible.

Multi-Camera Integration: Many AI video monitoring systems support multiple camera views. If an object is occluded in one camera, another camera from a different angle may still capture it, helping maintain situational awareness.

Scene Context and Predictive Modeling: AI models analyse the historical and environmental context, allowing them to anticipate where an occluded object is likely to reappear. This is especially useful in perimeter protection and behavioural analytics.

Edge and Cloud-Based AI Processing: AI systems use advanced edge computing and cloud-based processing to continuously refine detection accuracy, reducing false alarms and improving object classification even when partial occlusions occur.

Adaptive AI Learning: Continuous AI training ensures the system adapts to scenarios where occlusions frequently occur, such as people walking behind parked cars or moving objects in a cluttered environment.

Noise-Cancelling and Scene Filtering: AI algorithms apply noise filtering techniques to enhance detection accuracy by distinguishing relevant and irrelevant movements. This is useful for reducing false alarms caused by temporary obstructions like foliage or reflections.

AI-Powered Anomaly Detection: Some AI video analytics platforms, like those used in false alarm filtering, detect anomalies by comparing current video feed data with baseline scene models. This helps identify partially hidden objects by analysing their behaviour over time.

While AI is highly advanced, complete occlusions can still pose challenges. However, AI-based systems leverage redundancy across multiple data points to ensure that security and surveillance operations remain effective.

AI improves over time through continuous learning, primarily via the following mechanisms:

Data-driven learning and Model Training

AI systems continuously ingest new data, allowing their models to learn and adapt. The more data they process, the better they become at identifying patterns, making predictions, and reducing errors. AI video analytics, for instance, use machine learning and deep learning to analyse video content, classify objects and detect events.

False Alarm Reduction and Accuracy Enhancements

AI video monitoring systems, improve by filtering out false alarms. This is achieved through advanced algorithms that refine their ability to distinguish between real threats and benign activities. Over time, these systems reduce false positives, which improves operational efficiency in security applications.

Continuous Model Updates and Retraining

AI models undergo regular updates based on new real-world data, refining their ability to handle complex environments.

Self-Learning Algorithms and Adaptive Features

Some AI systems have built-in adaptive learning mechanisms. For example, AI can be reconfigured based on prior learning to improve recognition accuracy. AI video monitoring solutions can continuously refine their object classification abilities by learning from past detections.

Integration with Cloud Computing for Scalable Improvements

Cloud-based AI systems leverage large-scale datasets and computational power to improve efficiency and accuracy without requiring local hardware upgrades.

Behavioral and Predictive Analytics

AI-based security systems utilise behavioural analysis to improve anomaly detection over time.

Human Feedback and Hybrid AI-Human Collaboration

AI video analytics integrate human feedback to refine decision-making. Operators verify alerts, and the system learns from corrections.
AI can optimise its learning through reinforcement learning, where correct predictions are reinforced while incorrect ones are adjusted.

Real-Time Threat Detection and Continuous Improvement

AI in video surveillance detects and classifies objects in real-time, refining its accuracy as it gathers more operational data.
By reducing the number of false alarms, operators can focus on real threats, increasing the system’s overall effectiveness.

Conclusion

AI improves through continuous data ingestion, model updates, real-time learning, and human feedback integration. AI-enhancement reduces false alarms, refining threat detection, and integrating cloud-based updates. Over time, these mechanisms contribute to more accurate, reliable, and intelligent AI applications.

AI Video Monitoring is highly effective in detecting threats but is not infallible. Its accuracy depends on various factors, including environmental conditions, system setup, and AI model training. Here are some key insights from different AI-powered video monitoring solutions:

Accuracy of Threat Detection: AI Video Monitoring uses object detection and classification to differentiate between humans, vehicles, and other objects in real time. Evolon, our cloud-based AI algorithm provider, integrates advanced AI models that continuously learn and improve over time, increasing the accuracy of distinguishing genuine threats from non-threatening movements.

Challenges & Limitations: AI Video Monitoring is not always practical for accurate outdoor detection due to environmental factors like automobile light reflections, rain, snow, blowing leaves, and small animals. These conditions contribute to false positives and, in some cases, may lead to missed detections.

<>Use of AI Filtering: before escalating an event to a human operator, AI filters the video to determine if the scene contains a person or a vehicle. AI first determines whether an alarm is a “true positive” or “false positive.” If it is a true positive, the system sends the event to an operator for further action.

Continuous AI Training: AI models used in video analytics are continually trained and updated to adapt to new scenarios, improving accuracy over time. Cloud-based AI offers the advantage of real-time updates without the need for local software or firmware upgrades.

Best Practices for High Accuracy: To maximise accuracy, AI Video Monitoring systems should:

• • Use high-resolution cameras with adequate lighting and a clear field of view.

• • For proper recognition, ensure that objects to be detected are at least 10% of the camera’s field of view.

• • Have a fixed high-speed internet connection to process video analytics efficiently.

• Risk of Missing True Positives: Despite its effectiveness, AI Video Monitoring can miss true positives, which means they may fail to detect an actual threat. This is one of the significant concerns with AI surveillance, and users should implement human verification and other security measures as a backup.

AI Video Monitoring is a powerful tool for enhanced security, reducing false alarms, and improving operational efficiency. While AI models have shown high accuracy rates, they are not foolproof. Environmental conditions, object detection limitations, and AI training quality all impact detection performance. Businesses should combine AI monitoring with human oversight and complementary security measures to maximise threat detection accuracy.

Object detection in AI Video Monitoring is a computer vision technique for identifying and locating objects within an image or video. It involves classifying objects into predefined categories and determining their positions, allowing for the recognition and spatial understanding of various objects in a scene.

Key Aspects of Object Detection:

Classification and Localisation:

Object detection classifies objects (e.g., identifying a car, person, or dog) and localises them within the video stream by drawing bounding boxes around detected objects.

Applications:

Object detection is widely used in various fields, such as autonomous driving, security surveillance, healthcare, and retail analytics.

Techniques:

Modern object detection techniques often use deep learning models, such as:

• YOLO (You Only Look Once) – A fast and efficient real-time object detection algorithm.
• SSD (Single Shot MultiBox Detector) – Optimised for speed and accuracy.
• Faster R-CNN – Known for high accuracy in detecting objects.

Challenges:

Dealing with varying object sizes, occlusions, and complex backgrounds can make detection difficult.
Environmental factors such as lighting conditions and camera angles also impact accuracy.

Object Detection Role in AI Video Monitoring:

Enhanced Security: AI-driven object detection helps identify intruders, unauthorised vehicles, and other security threats.

False Alarm Reduction: AI filters out non-threatening movements such as shadows, animals, and weather conditions.

Proactive Monitoring: AI-powered analytics continuously analyse video feeds, providing real-time alerts for security personnel to take immediate action.

AI Video Monitoring incorporates several advanced artificial intelligence technologies to enhance security, reduce false alarms, and improve efficiency. The key technologies used include:

AI Video Analytics: Machine learning and deep learning algorithms analyse video to detect, classify, and track objects or events within a scene. This allows for automated video analysis beyond simple motion detection.

Object Detection and Classification: AI-powered object detection identifies and classifies objects like people, vehicles, and animals. AI-powered object detection identifies and classifies objects like people, vehicles, and animals. Convolutional Neural Networks (CNNs) like YOLO (You Only Look Once) and Faster R-CNN are often used.

False Alarm Filtering: AI filters out false alarms caused by environmental factors such as reflections, moving leaves, animals, and weather conditions. This significantly reduces the number of false positives operators need to handle.

Behavioural Analysis: AI is used to recognise and interpret human behaviour in video, identifying suspicious activities such as loitering, running, or other anomalies that may indicate security threats.

Scene Noise-Canceling: AI algorithms help eliminate scene noise, such as shadows and minor environmental changes, to prevent unnecessary false alarms.

Cloud-Based Video Analytics: Cloud-based AI processes video streams remotely, eliminating the need for local hardware and enabling continuous AI model updates for improved accuracy over time.

Real-Time Threat Detection: AI-powered systems continuously analyse video feeds, offering proactive monitoring to detect and prevent potential security threats before they escalate.

Perimeter Intrusion Detection: AI extends the range of video surveillance, using algorithms to detect and track potential threats over long distances. It integrates with PTZ cameras to zoom in and classify objects as people or vehicles.

Continuous AI Training and Improvement: AI models are continuously trained using real-world data to improve accuracy in detecting true threats and reducing false positives over time.

These AI-driven capabilities significantly enhance video surveillance operations by automating threat detection, reducing operator fatigue, and providing valuable security insights.

How can the combination of AI video surveillance and VMS enhance perimeter security?

AI-powered video surveillance and Video Management System (VMS) integrations can significantly enhance perimeter security by improving detection accuracy, reducing false alarms, and enabling real-time threat response. Here are key ways they contribute to better perimeter security:

Advanced Threat Detection & Reduced False Alarms

AI-driven surveillance systems use machine learning and computer vision to distinguish between real threats (e.g., intruders, vehicles) and non-threatening movements (e.g., animals, swaying trees). AI reduces false alarms, allowing security teams to focus on genuine threats instead of responding to unnecessary alerts.

Automated Real-time Monitoring & Response

AI-integrated VMS can automatically detect, classify, and track objects across multiple camera feeds. Real-time alerts enable faster response to security breaches, unauthorised access, or suspicious behaviour.

Perimeter Intrusion Detection & Smart Analytics

AI-powered cameras detect movement along a perimeter fence, track intruders, and trigger alarms.
Intelligent analytics can identify unusual behaviours, such as loitering or individuals attempting to bypass barriers.

Integration with Access Control & Security Systems

AI-driven VMS can be integrated with access control systems to trigger automated lockdowns when a breach is detected. AI enhances biometric authentication methods, such as facial recognition, to improve security at entry points.

Multi-sensor Fusion & Thermal Imaging

AI can combine data from multiple sensors (thermal cameras, LiDAR, and radar) to improve detection in low-light or adverse weather conditions. Thermal imaging AI analytics can identify unauthorised movement even in complete darkness.

Predictive Analytics & Proactive Threat Prevention

AI analyses historical security data to predict potential threats based on patterns of suspicious activity. Predictive analytics can suggest adjustments to security protocols based on risk assessments.

AI-powered Video Search & Forensics

AI-integrated VMS enables rapid search through video footage using attributes such as clothing colour, vehicle type, or facial recognition. Security teams can quickly investigate incidents without manually reviewing hours of footage.

Remote Monitoring & Cloud-based Security

AI-enabled VMS allows security teams to monitor and manage surveillance remotely from any location. Cloud-based integrations enable centralised security management for multiple sites.

Automated Drone & Robotics Integration

AI-powered drones and robotic security patrols can be deployed automatically to investigate perimeter breaches. These systems provide an additional layer of security for large or difficult-to-monitor areas.

Compliance & Incident Documentation

AI-enhanced VMS helps maintain compliance with security policies by automatically logging incidents and generating reports. AI-powered auditing tools help ensure that security teams adhere to best practices.

Conclusion

Integrating AI with video surveillance and VMS enhances perimeter security by providing proactive threat detection, reducing false alarms, and enabling automated responses. These advancements help security teams improve efficiency, reduce operational costs, and ensure higher protection for critical infrastructure, commercial properties, and high-risk areas.

AI Video Monitoring is an advanced video surveillance service integrating Artificial Intelligence (AI) with a video surveillance system to enhance monitoring accuracy, reduce false alarms, and automate threat detection.

How AI Video Monitoring Works

Event Detection: The customer’s video surveillance system detects an event and sends video to the monitoring platform.

Video Retrieval: The monitoring platform retrieves the associated video clip for analysis.

AI-Based Alarm Filtering: An AI algorithm analyses the video and evaluates whether the alarm is a “True Positive” (genuine threat) or a “False Positive” (non-threatening event).

Operator Action: If the AI determines the alarm is a true positive, an alert is sent to a monitoring centre for further action.

Key Components
Uses machine learning to detect, classify, and track objects such as people, vehicles, or anomalies in video streams.

False Alarm Reduction: AI helps filter out false alarms caused by environmental factors (e.g., shadows, animals, leaves) to improve monitoring efficiency.

Object Detection & Classification: AI models identify specific objects in video streams, distinguishing between humans, vehicles, and other elements.

Benefits of AI Video Monitoring

Reduced False Alarms: Up to 95%* fewer false alarms, allowing security teams to focus on real threats.

Improved Operator Efficiency: Improved Operator Efficiency: AI verification helps security teams avoid alert fatigue by only presenting verified alerts.

Scalability: Cloud-based AI systems allow easy integration with multiple cameras and monitoring sites without requiring extensive hardware.

Proactive Threat Detection: AI continuously learns from video data, improving accuracy over time.

Cost Savings: Reducing the need for physical guards and unnecessary alarm responses. Improves compliance with Right to Disconnect legislation.

Challenges & Considerations

Potential Missed Alarms: AI-based systems are not infallible and may sometimes fail to detect genuine threats.

Environmental Limitations: AI effectiveness depends on lighting conditions, camera resolution, and object size in the field of view.

Data Privacy & Compliance: AI video analytics must comply with data protection regulations when processing video.

Common Use Cases

Perimeter Protection: Continuous monitoring of secured outdoor areas to detect intrusions.

Manned Guarding Replacement: AI-driven monitoring for security operations that traditionally relied on on-site guards.

Comply with Right to Disconnect legislation: Improve compliance by stopping staff from being woken by calls about false alarms.

AI Video Monitoring refers to integrating Artificial Intelligence (AI) with video surveillance systems to enhance the effectiveness and efficiency of monitoring tasks. It employs advanced AI algorithms, such as machine learning and deep learning, to analyse real-time video streams for specific events, objects, or behaviours.

Our monitoring centre operates 24/7, and viewing is event-triggered — operators review the video around each activation rather than staring at a wall of idle screens. That’s a feature, not a compromise: it’s what makes monitoring accurate, affordable and scalable across every camera you own.

The industry term for exactly this service: using footage to confirm whether an activation is a real event before anyone responds. Verified events carry weight — with responders, with insurers, and in your own records.

It’s priced monthly and depends on camera count, the trigger mode (detector or analytic), and your response plan. Detector-triggered monitoring also pairs with a compatible [alarm monitoring] service. Request a consultation, and we’ll scope it against your actual site.

In most cases, yes — professional IP camera systems are generally suitable. We assess your system, then commission and prove every camera view before live monitoring begins.

Operators work with pre-set views mapped at commissioning — they assess what each activation shows, rather than driving cameras around your site. Critical coverage areas are agreed up front so the right views exist before they’re needed.

The service follows your armed state and the coverage windows agreed at commissioning — most sites run it after hours and while unoccupied.