Hi! We are Carlota Alcubilla and Nerea Alonso, and we have been working for the last few months with an internship at Quasar on Vessel Detection using Multispectral Images from the Sentinel-2 Satellite mission, specifically focusing on the southwest region of Mallorca.
During our internship, our main task was to detect boats using Sentinel-2 images and compare them with data from an AIS (Automatic Identification System) data website. Our workflow began by masking the land areas in the Sentinel images, which allowed us to focus solely on the water. Then, we applied the CFAR (Constant False Alarm Rate) method to detect the vessels in the sea. After detecting the boats, we saved their coordinates along with an image where the boats are classified by their area, as we can see in the image below.
On the other hand, we also georeferenced the captures from the AIS data website to extract the coordinates of the detected vessels. This allowed us to later compare these AIS-detected boats with those identified in the Sentinel-2 images. The following images display the ships detected in the AIS capture and a comparison between the ships identified using Sentinel data and those found with AIS data.
During our time working on this project, we have learned a lot—not only have we improved our image processing and coding skills, but we have also gained insight into how satellite data can be applied in real-world scenarios. Additionally, we have become familiar with various data formats and georeferencing techniques.
Overall, our internship on this project has been an enriching experience, providing us with valuable knowledge and the opportunity to work on a significant, real-world problem from start to finish. The supportive and nurturing environment at Quasar made this journey truly remarkable. We are grateful to Quasar for this incredible opportunity.