VarroDetector

VarroDetector is a free and open-source tool designed to identify and count Varroa mites in images of sticky sheets taken with smartphones. The software runs in low-range computers (no GPU is needed). Neither installation nor internet connection is needed, just double-click on the executable.

The detection process is based on a YOLO nano model (currently, YOLOv11 nano) specifically trained on hundreds of images to identify Varroa mites.

🟢 🐝 Benefits

The well-known beekeeper The Apiarist has written a post about the program on his popular blog, highlighting its ease of use and showing how it works, with comments such as:

• "I'm sure there will be (and maybe already are) other AI applications that benefit bees or beekeepers, but this is the first that I've seen that is both practical and relevant."
• "I've only analysed a few of my own images so far, but it’s far more impressive than anything I've seen before. It's easy and quick to load and automatically process the images, then scan them visually for any glaring omissions or inclusions."

⚡ Execution Options

Option 1: Executable File (Recommended)

1. Download the VarroDetector executable file for Windows or for Linux (MacOS coming soon).
2. Run the executable file

Note: This file is self-contained, so the application takes a few seconds to start since the contents must be unzipped on the fly.

Option 2: Run from Source

1. Clone this repository

git clone https://github.com/jodivaso/VarroDetector.git

2. Install the required dependencies:

pip install -r requirements.txt

3. Run the application:

python varroa_mite_gui.py

📖 Citation

If you use this software, please cite this paper:

Yániz, J., Casalongue, M., Martinez-de-Pison, F. J., Silvestre, M. A., Consortium, B., Santolaria, P., & Divasón, J. (2025). An AI-Based Open-Source Software for Varroa Mite Fall Analysis in Honeybee Colonies. Agriculture, 15(9), 969. https://doi.org/10.3390/agriculture15090969

Bibtex:

@article{VarroDetector,
  title = {An AI-Based Open-Source Software for Varroa Mite Fall Analysis in Honeybee Colonies},
  volume = {15},
  ISSN = {2077-0472},
  url = {http://dx.doi.org/10.3390/agriculture15090969},
  DOI = {10.3390/agriculture15090969},
  number = {9},
  journal = {Agriculture},
  publisher = {MDPI AG},
  author = {Yániz,  Jesús and Casalongue,  Matías and Martinez-de-Pison, Francisco Javier and Silvestre,
Miguel Angel and Consortium,  Beeguards and Santolaria,  Pilar and Divasón,  Jose},
  year = {2025}
}

Features

• String Detection: Automatically identifies and crops images based on guide strings
• AI-powered Detection: Uses YOLO nano learning model to identify Varroa mites
• Confidence Threshold: Adjustable detection sensitivity per image or globally
• Region of Interest (ROI): Define specific areas for mite counting
• Subfolder Support: Processes nested folder structures
• Comprehensive Statistics: Per-image, subfolder, and total counts
• Raw File Support: Processes both JPG and DNG camera files

User's manual

The program is very easy to use; however, you can click here to download the user's manual in PDF format.

Controls

• Zoom: Mouse wheel
• Pan: Middle mouse button
• Add mite manually: Left click and drag
• Delete mite: Right click on detection box
• Hide detections: Press and hold 'h' key
• View different images: Click on image names in the list

Working with Regions of Interest (ROI)

1. Select an image
2. Click "Edit ROI"
3. Left click to add points around your area of interest
4. Double click to complete the polygon (will be drawn in yellow)
5. Statistics will update to count only mites within the ROI
6. Right click to delete the current ROI

Saving Results

When you click "Save Results", a new folder named "results" will be created in your input folder containing:

• images/: All processed images with visible detection boxes
• labels/: YOLO format text files with detection coordinates
• roi.txt: A text file containing the points of the Region of Interest of the images
• statistics.csv: Detailed counts for each image
• statistics_subfolders.csv: Summary statistics by subfolder

Output Format

statistics.csv

• filename: Relative path to the image
• threshold: Confidence threshold used for detection
• varroa_count: Number of mites detected in this image
• subfolder_count: Total mites in this image's subfolder
• total_count: Total mites across all images

statistics_subfolders.csv

• folder_name: Subfolder path
• num_varroa_mites_folder: Total mites detected in this subfolder
• threshold: Confidence threshold(s) used
• num_images: Number of images in this subfolder
• name_images: List of image filenames

Sample images

Some sample images to test the program can be downloaded from here.

Acknowledgments

This research has been funded by:

• Grant INICIA2023/02 by La Rioja Government (Spain)
• MCIU/AEI/10.13039/501100011033 (grant PID2023-148475OB-I00)
• The EU Horizon Europe (grant 101082073)
• The DGA-FSE (grant A07_23R)

License

This software uses a YOLOv11 nano model; thus, it is licensed under the GNU Affero General Public License v3.0 (AGPL-3.0).