Selected market overview of AI based products for “Situation 
Awareness” 
There are currently no independent studies and comparisons between current devices available on the 
market. A review of most of the products currently offered, containing Al components, is based on the 
corresponding publicly available product descriptions (Koch et al., 2024). It must be taken into consid- 
eration that the product descriptions of the manufacturers might be biased and lack some necessary 
information because it acts as marketing material. One extract from this study (Koch et al., 2024), in- 
cluding night vision systems, is summarised in Table 1. It shows that 10 out of the 11 products re- 
viewed, which include night vision systems, use additional daylight cameras to support the nautical 
personnel with the task of object identification. An inertia measurement unit (IMU) sensor is used in 
five out of the 11 products reviewed. The review highlights the theoretical importance of EO systems 
in the maritime environment. Maritime AIS is regulated by IEC 61993 and IEC 62287. 
Data obtained from EO sensor systems has a decisive advantage in obstacle and ship detection over 
data obtained from AIS and Radar. When interpreted correctly, Al-supported systems can notify per- 
sonnel about unforeseen obstacles such as floating containers or unmapped land masses. If the sys- 
tems receive only AIS or Radar data, they can detect only obstacles equipped with appropriate and 
activated transmitters and obstacles of a certain size. If the technology improves further and makes IR 
cameras more accessible, the EO sensor system could become a core sensor in autonomous vessels. 
(Thombre et al., 2022)(IMO, 2024). 
' RGB Camera 
Night vision! LIDAR/RADAR | IMU_ | AIS * 
u 
tx 
Avikus HiNAS 2.0 
Avikus NeuBoat ' 
Awarion 
Awarion 
MTU NautlQ Copilot 
Orca Al 
Robopec 
Sea.Al 
Seasight 
' Wärtsilä Voyage Autonomy Solutions . 
Yara Birkeland X a 
Table 1: Comparison of selected Al based systems in the 2023 market survev (Koch et al., 2024) 
N 
WE 
X 
X 
xt 
X 
* 
X —m—— 
X X 
X X 
yo rw 
Segmentation and object detection 
The study on “Evaluation of Algorithms for Nautical Applications” (Koch et al., 2024) highlights different 
aspects focusing on the standardisation and regulation of Al-based systems used in EO sensors. Object 
detection enables EO systems to classify different objects. In a nautical context, these include ships, 
Duoys, obstacles, and navigation aids as well as harbour infrastructure. The exact detection of these 
objects is fundamental for ensuring safe navigation and collision avoidance as well as monitoring ship 
traffic in ports. The study (Koch et al., 2024) compares different Machine Learning architectures, taking 
into consideration the challenges of the maritime environment, and benchmarks its performance in 
general situations. To demonstrate the principle, this study focusses on the detection of ships. In 2023, 
no sufficiently comprehensive and annotated datasets were published for other nautical objects such 
as buoys, obstacles, or navigation aids. Two types of applications were subject to the investigation: