Varroa Destructor Classification Using Legendre-Fourier Moments with Different Color Spaces

Area of Science:

• Agricultural Entomology
• Computer Vision
• Biotechnology

Background:

• Bees are vital for global food production through pollination.
• The Varroa destructor mite is a major threat, causing viral infections and hive collapse.
• Early disease detection is essential for bee colony health and preservation.

Purpose of the Study:

• To develop an innovative method for early detection of Varroa destructor mites in honey bees.
• To evaluate the efficiency of multichannel Legendre-Fourier moments for mite identification.
• To compare the proposed method with existing deep learning techniques.

Main Methods:

• Utilized multichannel Legendre-Fourier moments for image analysis of honey bees.
• Employed a subdivided VarroaDataset to enhance feature extraction (color, shape of bee body parts).
• Compared the proposed algorithm against DeepLabV3 and YOLOv5 for semantic segmentation and object detection.

Main Results:

• The Legendre-Fourier moments approach demonstrated effective identification of Varroa destructor mites.
• The method showed robustness with rotation and scale invariance, and noise resistance.
• The proposed technique offers a promising alternative to current deep learning methods for mite detection.

Conclusions:

• The developed method provides a novel and effective tool for early Varroa destructor mite detection.
• This technology can significantly contribute to bee preservation strategies.
• Protecting bee populations is critical for maintaining agricultural productivity and food security.