Improving the Accuracy of Habitat Distribution Estimation of Scopimera Globosa on Tidal Flats Using Machine Learning

Key Points

• The study aims to enhance the accuracy of estimating habitat distribution of Scopimera globosa using machine learning.
• Applied machine learning for automatic detection of Scopimera globosa burrows.
• Implemented data augmentation on training data to improve detection accuracy.
• Assessed precision and recall metrics for model performance.
• Achieved a precision of 0.676 and recall of 0.736 in detecting crab burrows.
• Detected seasonal variations in crab density analogous to previous field observations.
• Demonstrated successful burrow detection across different environmental conditions.