Passenger ship

vision for

  • Maritime archaeology
  • Search and rescue
  • Fishing and biology
  • Underwater structure maintenance
Become a patron

About us

At MaritimeAI we research and develop all kinds of computer vision applications for social good (non-commercial) and business domains, which aim to help search & rescue operations, maritime archaeology and biology, leisure diving activity, and offshore oil & gas and communications operations.

Video and Image Processing

As a computer vision group, we work with video and image processing, mainly for underwater applications.

Underwater Imagery Dehaze and Color Restoration
existing use case study

Initially, underwater surveys with unmanned underwater vehicles had their major tradeoffs with the technical specifications. Tradeoffs wich led to various flaws, such as low-quality video capturing and low degree of control over it. Today, one of the bottleneck links in the chain of surveys is a man behind the unmanned underwater vehicle (UUV), who operates it and trying to get an adequate picture of what's being viewed on the screen.

Low quality of the content can be explained by several facts: cameras are too cheap, too much mud or organic particles, not enough ceiling light.

We can propose a solution to the case, given the fact that research on parts of the problem has been made. Our turn is to make further research and to bring it to the customers polished, fast, and beautiful to make the work of UUV operators simpler and surveys cheaper.

To achieve it, we are developing our approach, which includes several traditional methods of processing the video with the usage of modern architectures based on deep neural networks.

Underwater Object Detection and Classification

We aim to help maritime biologists to monitor sea life changes, fish migrations, coral colonies evolution as well as any hazards present for life under the sea.

We also use our computer vision expertise to identify the types of objects on the seabed and drowned bodies on sonar images.

Fish Families Detection
existing use case study

For a long time in a commercial fishing industry bycatch and overfishing were paramount problems. A quarter of the world's captured fish ends up in throwing overboard because it's being caught unintentionally – not only unwanted species or inferior quality of fish but also sea turtles, sea lions, and other mammals.

A lot of effort was made to catch what was intended to catch and release the fish that is of no use. We lead research in collaboration with maritime scientific institutions and industry equipment producers to make the industry both ecology friendly and more effective.

We apply modern deep neural network technologies to underwater data analysis. This is intended to help ships see, what they catch, minimize damage to nature, and make industrial fishing effective at the same time.

Drowning Victims Identification
existing use case study

Normally, search, rescue, and recovery (SRR) operations take a lot of time and effort as operators have to search manually across the area of investigation, constantly looking at sonar screens for 12-14 hours in a row or trying to identify missing ships on satellite or drone images.

We aim to use both our knowledge in computer vision and domain expertise to help SRR teams discover the aftermath of an accident within a smaller time frame and with greater precision.

Our technologies are used by some of our partners to find drowning victims in inland waters, both lakes and rivers, decreasing time-to-discovery from days to hours.

Object Detection and Classification
existing use case study

One of our key research directions implements the automatic detection of various objects on the seabed surface. Types of objects range from large ships, airplanes, submarines to small boats, cars, lost equipment, even human bodies. This application addresses several important problems, such as maritime ecology, accident investigation, historical and archaeological research. At the same time, it can be essential for search, rescue, and recovery operations, where a fast and precise scan of the seabed is required.

Fast and precise object detection on side-looking sonar images is also crucially important in surveys, conducted for the construction of underwater pipelines, cables, and other subsea facilities at all stages: from searching for the location of the facility to inspection and decommissioning.

The implemented technology uses pure sonar images in combination with state-of-the-art machine learning and deep learning approaches to deliver the detection results. Thus providing great automation opportunities for all the kinds of underwater search at large scale, and a significant boost for the following data analysis.

Surface Object Detection and Classification

Surface object detection and classification are one of the main aspects of maritime computer vision. Using artificial intelligence techniques, we build object (life rafts) detection on Synthetic aperture radar and satellite images, identify unlit vessels in the IR spectrum.

Ice Identification and Classification
existing use case study

Arctic ice has always been a great danger for naval operations in the North. However, Nordic countries, including Russia, increase the number of ship routes in the northern seas, mainly for natural gas transportation purposes.

At the same time, ice breakers fleet has limited capacity, thus making shipping operations costly, as each convoy route planning takes a lot of time and effort to avoid dangerous zones with high-density ice and ice ridges.

We have developed a demo case for our partner, Marine Research Center, to reduce the time of ice situation analysis to provide faster route planning and use ice breakers fleet more effectively.

Consulting on Specific Maritime Use Cases

We also provide consulting on specific use cases in maritime research and production applications.

Maintenance of Underwater Structures
proposal use case

We are researching towards automated computer vision solutions using sonar imagery (installed on AUVs or man controlled min-subs) which make track of deformations, leakages or any other potentially hazardous situations easier.

Another possible application is to make a preliminary analysis of the seabed structure before the installation of any structures or equipment.

Maritime Archaeology
proposal use case

Maritime archaeology has been very challenging in terms of actual search operations. We are working on solutions that help to identify anomalies and man-made structures on the seabed, rank them by the respective probability to be an actual point of interest.

Secondly, we apply classification algorithms to identify the type of wreck and provide accurate information about the object.

Thirdly, we help to create a detailed view of the artifact using computer vision-based segmentation.

Last, but not the least, we use 3D modeling and Generative Adversarial Networks to create artificial datasets in situations, where real-world data is unobtainable otherwise.


We are looking for collaboration with maritime agencies, research institutes, maritime archeology entities, as well as investors and maritime equipment producers and resellers.