Abstract
Recently, in order to help officers on watch and vessel traffic service officers have a quick view of the maritime traffic situation, a map of the distribution of places with a high likelihood of collisions is a new and promising approach. In this study, the authors will apply a ship domain model to identify areas with high collision risk, thereby building a traffic hot spot map. The hot spot map is built in 3 steps: Collecting and processing AIS data in Vietnam's seas; Apply the DBSCAN clustering method to group ships that are close together into clusters with at least 2 ships or more; Ship domains will be applied to ships in the same cluster to determine where collisions are likely to occur. Recorded traffic hotspots are areas where ship domain overlaps with each other, creating a heat map in Vietnam waters. The effectiveness of the model was tested in the Hai Phong area, giving intuitive results, hot spots are shown as colors combined with specific values, which can be quickly identified and evaluated.