Military Aircraft Detection

Annotated image dataset for the automatic detection of 43 types of military aircraft, with PASCAL VOC enclosing boxes.

Download dataset

Size

4,388 annotated images, VOC format, 624 MB

Licence

Apache 2.0

Description

‍

The dataset Military Aircraft Detection contains 4,388 annotated images for the detection of military aircraft using enclosing boxes in PASCAL VOC format. It covers 43 different models (F-16, Rafale, B-2, Mirage2000, etc.) and is a valuable resource for training or testing object detection algorithms in a military or aeronautical context.

‍

What is this dataset for?

‍

Train object detection models like YOLO, SSD, or Faster R-CNN on military aircraft
Develop visual recognition systems for military use or aerial surveillance
Perform aircraft classification and counting on aerial or satellite images

‍

Can it be enriched or improved?

‍

Yes, we can improve this dataset by adding additional annotations (orientation, estimated altitude, weather conditions), by diversifying the sources (infrared, satellite), or by standardizing class names. It is also possible to supplement it with synthetic images to reinforce low-volume training.

‍

🔎 In summary

Criterion	Evaluation
🧩 Ease of use	⭐⭐⭐⭐✩ (Clear structure and standardized annotations)
🧼 Need for cleaning	⭐⭐⭐⭐✩ (Low – check for possible duplicates)
🏷️ Annotation richness	⭐⭐⭐✩✩ (Bounding boxes only, no segmentation)
📜 Commercial license	✅ Yes (Apache 2.0)
👨‍💻 Beginner friendly	✅ Yes – easy to use with a standard framework
🔁 Fine-tuning ready	✅ Excellent base for fine-tuning a detection model
🌍 Cultural diversity	🌎 Medium – airplane models mainly American, Russian, European