Ships in Aerial Images
Dataset composed of 26,900 annotated aerial images specifically for ship detection. Annotations are provided in YOLO format, ensuring accuracy and efficiency. The dataset makes it possible to address many issues related to maritime safety and resource management.
Description
The “Ships in Aerial Images” dataset contains a vast collection of annotated aerial images to detect unique ships. The annotations are in the form of surrounding boxes in YOLO format, adapted to modern object detection algorithms.
What is this dataset for?
- Develop and train ship detection models in aerial imagery
- Maritime surveillance for safety, fisheries management, and environmental protection
- Applications in defense, anti-piracy and maritime border control
Can it be enriched or improved?
It is possible to enrich this dataset by adding additional classes (ship types) or more precise annotations (segmentations). Temporal data could also be integrated for video detection.
🔎 In summary
🧠 Recommended for
- Computer vision researchers
- Maritime surveillance professionals
- AI developers
🔧 Compatible tools
- YOLO
- TensorFlow
- PyTorch
- OpenCV
- Detectron2
💡 Tip
Use YOLO annotations directly with popular frameworks for a quick start.
Frequently Asked Questions
What is the main annotation used in this dataset?
The dataset uses surrounding boxes in YOLO format to annotate the presence of ships.
Can this dataset be used for real-time detection?
Yes, thanks to the YOLO format and the size of the dataset, it is adapted to real-time applications.
Does the dataset include images with different types of ships?
No, the dataset only contains a “ship” class without distinction of types.
