Frequently asked questions

SEA.AI can detect any floating object on the water surface, regardless of the material : wood, steel, plastic, organic matter, or ice.

SEA.AI detects objects that conventional systems like radar or AIS might miss, providing an additional layer of safety to your vessel and crew.

The highly sensitive thermal cameras ensure reliable detection day and night.

SEA.AI classifies objects such as sailing and power boats, ships, buoys, persons in the water (MOB – Man Overboard), whales, maritime mammals, and icebergs. Any unknown objects are marked as an unknown hazard, ensuring no potential threat goes unnoticed at sea.

SEA.AI combines advanced thermal and color cameras with state-of-the-art AI to deliver reliable detection in any environment.

Whether at night, in low visibility, or during challenging weather, SEA.AI continuously monitors the surroundings, detecting hazards that are invisible to the human eye or traditional cameras.

No matter the task, navigation, steering, or other duties, SEA.AI acts like a vigilant “lookout” atop the mast, instantly alerting the crew when a potential obstacle or danger appears.

Performance in various conditions:

• Light rain or fog:
  

  The system remains fully operational. While detection range may be slightly reduced due to atmospheric attenuation, SEA.AI continues to provide early and reliable alerts. You can see SEA.AI operating in real fog conditions in this live demonstration video.

• Heavy rain, thick fog, or high waves: Detection may be somewhat reduced, but SEA.AI continues to provide reliable alerts.
• Night or low visibility: Thermal imaging ensures hazards are detected even when human vision is limited.

With SEA.AI, crews gain constant vigilance and enhanced safety in all conditions.

Day or night, calm seas or challenging weather, SEA.AI provides peace of mind and confidence for recreational sailors, commercial operators, and security teams alike.

SEA.AI, Radar, and AIS are complementary maritime safety technologies, but they operate on fundamentally different principles and serve different detection purposes across recreational, security, and commercial applications.

Radar is an active detection system that emits radio waves to identify objects around a vessel. It is widely used on commercial ships, patrol and security vessels, and many recreational boats. Radar performs reliably in reduced visibility and is effective at detecting larger targets such as ships and coastlines. However, it may struggle to detect small, low-profile, or weakly reflective objects on the water surface.

AIS (Automatic Identification System) is an active transponder-based system primarily mandated for commercial vessels and commonly installed on larger recreational and security boats. It broadcasts a vessel’s identity, position, speed, and heading to surrounding traffic. AIS improves situational awareness in busy waterways and shipping lanes, but it only detects cooperative targets equipped with an active transponder. Objects without AIS, including small crafts, floating debris, wildlife, or persons overboard, remain invisible to the system.

SEA.AI is a machine vision system based on thermal (infrared) and optical color cameras. Using the latest artificial intelligence technology, it analyzes live video data to detect anything floating on the water surface, regardless of radar signature or AIS transmission status. SEA.AI can identify objects even if they are not equipped with AIS or if their transponder is inactive.

Designed to be intuitive and easy to use, SEA.AI provides an additional layer of safety by detecting and alerting crews to objects that radar or AIS may miss. This added capability enhances situational awareness for recreational sailors, supports threat and surface monitoring for security operators, and strengthens collision avoidance procedures for commercial vessels.

Rather than replacing Radar or AIS, SEA.AI complements them, combining cooperative tracking, radio detection, and AI-powered visual recognition to create a more complete picture of the vessel’s surroundings.

Unlike a standalone thermal camera, SEA.AI goes far beyond simply displaying a thermal image. While thermal cameras are excellent tools for visibility in low-light or night conditions, they require constant human monitoring to detect and interpret potential threats — leaving room for human error, especially when you’re occupied with other tasks at the helm.

SEA.AI seamlessly integrates thermal cameras with cutting-edge AI and computer vision technology, transforming a passive visual tool into an active collision avoidance system. The AI automatically detects and tracks floating objects in the maritime environment in real time, issuing audible alarms the moment a potential risk or hazard is identified — even when you’re not actively watching the screen.

This is the key difference: a thermal camera shows you what’s there, but SEA.AI understands what it sees. By combining thermal imaging with machine vision, SEA.AI delivers AI-driven object detection, continuous perimeter surveillance, and intelligent collision avoidance — giving sailors and mariners a critical safety advantage that no traditional thermal camera can offer on its own.

For anyone seriously considering a maritime safety solution, SEA.AI doesn’t replace your thermal camera — it makes it dramatically smarter.

SEA.AI systems can see up to the horizon, but the actual detection range is dependent on the mounting height, the size of the object above the water, the SEA.AI model, the environmental conditions and distance.

SEA.AI systems are in example able to detect :

– Sailing Boat up to 7500 m with the Sentry, for Watchkeeper the distance can go up to 2500 m depending on the model

– Man Over Board up to 700 m with the Sentry, 220 m maximum for Watchkeeper according to the model

Currently, the SEA.AI application supports up to two simultaneous user sessions, allowing crew members to monitor and interact with the system from two devices at the same time, whether from a navstation, a tablet, or a smartphone on board.

Additionally, SEA.AI allows multiple RTSP clients to connect to the live video feed simultaneously, making it easy to integrate with existing onboard display systems. The number of concurrent RTSP connections is only limited by the available network bandwidth on your vessel.

This flexibility ensures that the whole crew stays informed and connected, no matter where they are on the boat.

Yes, SEA.AI is highly effective during port manoeuvres. The system acts as an extra pair of eyes — especially useful at night or when visibility is limited (e.g. behind jetties or pontoons). Mounted at a high point on the vessel, SEA.AI provides a wider field of view than the human eye at deck level, helping to anticipate potential collisions with other moving vessels.

Watch a real-world testimonial on YouTube

SEA.AI is specifically designed to detect and classify objects that are on or slightly above the surface of the water — where the greatest collision risks occur. Using thermal and optical cameras powered by AI, the system reliably identifies floating debris, marine mammals, containers, and people in the water, both day and night.

While SEA.AI does not detect fully submerged features like reefs or sandbanks, it is built to complement traditional navigation tools such as sonar, radar, and depth sounders — extending a vessel’s awareness where human vision and conventional sensors may fall short.

By focusing on surface-level hazards, SEA.AI fills a critical gap in maritime safety — especially in low-visibility environments like night-time approaches, crowded ports, or open-sea drift zones.

Yes, SEA.AI can detect marine mammals — including whales — when they surface to breathe. This is a critical moment for collision risk, especially in regions where shipping lanes intersect with migratory routes.

Our AI processes multiple frames per second and identifies floating objects, including marine mammals, when even a small part of them is visible above the water surface. To improve this capability, we are continuously training our detection models using dedicated whale datasets, with the goal of contributing to marine life protection.

Learn more about our commitment to whale safety

Yes. SEA.AI uses long-wave infrared (LWIR) thermal cameras, which can detect objects even if they are at the same temperature as the surrounding water. This is possible due to differences in emissivity, reflectivity, and infrared radiation from different materials.

Each material emits heat differently. For example, water has an emissivity of ~0.94, while human skin is ~0.98. Even a small difference like this creates a clear infrared contrast detectable by our system. A metallic container and seawater may be the same temperature, but they emit different radiation patterns — making the object visible in thermal imaging.

SEA.AI uses uncooled LWIR technology rather than the more export-restricted MWIR (Mid-Wave Infrared) systems. LWIR cameras are better suited for lower temperature ranges (<100°C), and are less affected by tropical humidity than MWIR, which are typically used in aerospace and defense (e.g., missile tracking).

While MWIR can offer higher sensitivity (≈15 mK vs. 30–50 mK for LWIR), our LWIR sensors are precise enough to distinguish small thermal differences and are highly effective for maritime detection use cases.

SEA.AI combines cutting-edge thermal and optical camera technology with advanced artificial intelligence to detect, classify, and track objects on the surface of the water — both day and night. It alerts the crew in real time about potential hazards such as buoys, small boats, floating containers, marine mammals, or people in the water.

Founded in 2018, SEA.AI (formerly known as OSCAR) has become a pioneer in AI-powered maritime vision systems. Our systems are used by leading yacht builders, coastguards, and maritime professionals worldwide — delivering proven performance in harsh conditions, including ocean racing, commercial shipping, and government fleets.

What sets SEA.AI apart is our AI’s ability to process multiple sensor feeds and identify hazards with high reliability, even in darkness, fog, or busy port environments — situations where radar and human vision may fall short.

SEA.AI is not just a product — it is a next-generation lookout system, trusted by the industry to enhance safety, situational awareness, and response time.

Machine Vision in maritime navigation refers to the use of advanced camera systems combined with Artificial Intelligence (AI) to detect, classify, and assess objects on the water surface in real time.

Unlike traditional marine sensors such as radar or AIS — which detect signals or movement — Machine Vision analyzes visual and thermal imagery to understand what an object actually is. Using deep learning and neural networks, the system can distinguish between:

• Small boats
• Floating debris
• Containers
• Marine mammals
• People in the water

This is especially valuable in situations where human vision is limited — such as at night, in glare, in busy harbors, or during offshore passages.

At SEA.AI, Machine Vision technology combines high-resolution optical and thermal cameras with continuously trained AI models specifically optimized for maritime environments. The system processes multiple frames per second, evaluates potential collision risks, and alerts the crew early and reliably.

In short, Machine Vision gives vessels a digital lookout — enhancing situational awareness and improving safety at sea.

At SEA.AI, we train our Machine Vision models using a large and continuously expanding database of annotated maritime images collected from real-world operations.

We apply supervised deep learning and backpropagation to train neural networks capable of accurately detecting and classifying vessels, floating debris, containers, and marine mammals.

To ensure reliable performance across diverse environments, we use data augmentation and systematic hyperparameter tuning, optimizing detection accuracy under low light, glare, and varying sea states.

Before deployment, each model is validated against separate test datasets to measure precision and real-world robustness.

At SEA.AI, we estimate distance using a fully passive, vision-based approach. Our system applies mathematical projection algorithms to map the two-dimensional camera image onto the real-world water surface.

Distance calculation is based on:

• Precise camera specifications
• The mounting height of the camera above sea level
• Horizon detection using Machine Vision
• Motion and orientation data from an integrated IMU (Inertial Measurement Unit)

By continuously determining the horizon line and vessel attitude, the system translates image coordinates into real-world distance estimates — without requiring active sensors such as radar or LiDAR.

This passive method allows SEA.AI systems to provide reliable distance estimation while remaining lightweight and energy-efficient.

Note: Distance estimation is currently available in SEA.AI systems with this functionality enabled. SEA.AI Brain does not include distance estimation at this time.

At SEA.AI, our collision avoidance system uses a proprietary risk assessment algorithm to evaluate every detected object in real time and determine its potential threat level.

When the system detects an object — such as another vessel, buoy, kayak, floating debris, or marine mammal — it analyzes three key parameters in the following order:

1. Object Size
The size of the detected object establishes its baseline relevance. Larger objects generally require greater safety distance and therefore represent a higher potential collision risk.

2. Distance
The system then evaluates how far the object is from the vessel. The closer the object, the higher its risk weighting, as shorter distances reduce reaction time and increase urgency.

3. Bearing & Projected Path
Finally, SEA.AI analyzes the object’s position relative to the vessel’s heading and projected trajectory. Objects directly ahead or intersecting with the vessel’s path are prioritized over those safely off to the side.

These three parameters are dynamically combined into a calculated threat level score, which determines whether and how the crew is alerted.

By continuously evaluating object size, distance, bearing, and motion context, SEA.AI delivers intelligent, context-aware warnings — helping crews focus on genuine risks while minimizing unnecessary alarms.