Advanced Maritime Vessel Detection in Satellite Imagery Using Masati-V2

Maritime vessel detection in satellite images address real time issues such as maritime security, illegal fishing and environmental monitoring, which require efficient tracking of vessels in large scale. Maritime vessel detection is challenged by limited annotated satellite datasets, varying sea conditions, and the presence of small or overlapping vessels. YOLOv10 is best for maritime vessel detection in satellite images due to its high accuracy, speed and efficiency in handling complex environments. It is able to detect small vessels against vast and cluttered backgrounds. YOLOv10 addresses these challenges through architectural features, which able to detect small objects like ships. It is scalable for tasks like illegal fishing detection and coastal security using satellite images. Using Masati-V2 dataset we have trained algorithms like YOLOv10s and YOLOv10m which are family of YOLOv10.Evaluation is performed using performance metrics which include precision and recall. According to our research, YOLOv10m model gives the best prediction for maritime vessel detection in satellite imagery.  Future research will concentrate on enhancing classification models to identify various vessel types, including smaller or overlapping ones, under diverse environmental conditions and also need to explore real-time vessel tracking using temporal data and predictive analytics can be explored to monitor vessels continuously across multiple satellite images for improved situational awareness.