Autofocus Entropy Repositioning Method Bioinspired in the Magnetic Field Memory of the Bees Applied to Pollination.

Area of Science:

• Robotics
• Bioinspiration
• Computer Vision

Background:

• Precision pollination requires advanced navigation and positioning systems.
• Existing methods often need extensive calibration and are sensitive to environmental conditions.

Purpose of the Study:

• To develop a bioinspired, unsupervised navigation method for precision pollination rovers.
• To enhance rover positioning accuracy and robustness using magnetic field memory principles.

Main Methods:

• Image segmentation in the HSV color system for real-time sharpness feature extraction using entropy and variance of the Laplacian.
• A complementary positioning method based on area feature extraction from active markers, analyzing color, noise, and vibrations.
• Utilizing bee magnetic field memory principles for navigation.

Main Results:

• The unsupervised method eliminates the need for prior calibration of target dimensions, histograms, and distances.
• The algorithm demonstrates robustness to local, temporal, and geophysical variations in magnetic declination.
• Effective operation is achieved even with small sharpness variations and without reliance on luminosity after scanning.

Conclusions:

• The proposed bioinspired method offers a robust and calibration-free solution for precision pollination rover navigation.
• The system's adaptability to environmental changes and minimal reliance on external factors present a significant advancement.