Automated Monitoring and Analysis of Avian Behavior and Hatching Success Using YOLOv8: A Case Study on Red-Wattled Lapwings

This study delves into avian behavior, particularly focusing on incubation practices and hatching success, which are crucial elements for bird conservation. An automated system is introduced., leveraging deep learning techniques., and is exemplified through its application to the Red-wattled Lapwing. The system integrates You Only Look Once version 8 (YOLOv8)., a cutting-edge object detection model., enabling precise identification of eggs., hatchlings., and adult birds. It achieves an impressive F1 score of 0.93 and a mean Average Precision (mAP) of 0.951. The approach involves processing video data to detect and categorize behaviors such as incubation., attention., unattended and absence., based on the spatial positioning of identified objects. Behavior duration is tracked to assess incubation effectiveness. Additionally., the system computes the hatching success rate by analyzing egg and hatchling numbers over time. Outlier data is excluded using the Interquartile Range (IQR) method., and the most consistent egg and median hatchling counts are utilized to ensure accuracy. The findings confirm that the system effectively monitors and classifies behaviors., providing a dependable method for calculating hatching success. Although centered on the Red-wattled Lapwing., this methodology can be adapted for other bird species., offering a valuable tool for ecological research and conservation efforts.