DefendIntelligence/vessel-detection

Vessel Detection

Gradio Space for detecting vessels in satellite imagery with a fine-tuned YOLOv8 model.

The main demo example is a multi-vessel satellite patch with 14 detections at the default confidence threshold.

Links

• Live Space: https://huggingface.co/spaces/DefendIntelligence/vessel-detection
• Model repository: https://huggingface.co/DefendIntelligence/vessel-detection
• Direct model download: https://huggingface.co/DefendIntelligence/vessel-detection/resolve/main/models/best.pt

Model

• Local file expected by the app: models/best.pt
• Checkpoint source: train-20260417T124314Z-fad9d3ed_best.pt
• Run source: infer-b88a2887
• Training name: super-visible-y8s-newlabels-focuslite-e45
• Family: YOLOv8s
• Main dataset: sentinel-2-rgb
• Local index mAP50: 0.7912

The GitHub repository does not store best.pt. Use the bootstrap command below and it will download the model from Hugging Face.

Run Locally

git clone https://github.com/anisayari/vessel-detection.git
cd vessel-detection
python run_local.py

Windows shortcut:

.\start.ps1

macOS/Linux shortcut:

bash start.sh

The script creates a local .venv, installs requirements.txt, downloads models/best.pt from Hugging Face, then starts Gradio at http://127.0.0.1:7860.

Useful options:

python run_local.py --download-only
python run_local.py --skip-install
python run_local.py --host 0.0.0.0 --port 7860

Use The App

1. Upload an RGB satellite image or select an example.
2. Adjust the confidence threshold if needed.
3. Click Detect vessels.

The app tiles large images before inference so small vessels remain visible to the model.

Hugging Face Deployment

git init
git lfs install
git remote add origin https://huggingface.co/spaces/DefendIntelligence/vessel-detection
git add .
git commit -m "Add YOLOv8 satellite boat detector Space"
git push -u origin main

If the Space already exists, clone it and copy this folder's contents to the Space repository root.