Advanced Computer Vision Methods for Avian Risk Assessment Using Multispectral Drones

The growing threat of the Avian Influenza virus has raised the alarm for conservationists, ecologists and researchers to develop innovative solutions to monitor wildlife birds using drones and early diagnosing through advanced image processing methods. In recent years, outbreaks of the Avian Influenza virus have devastated avian populations worldwide, leading to significant declines in bird colonies, particularly in vulnerable species such as the Dalmatian Pelicans (Pelecanus crispus). These significant changes to the environment may contribute to a Cascade effect impacting the food chain of both predators and preys. This study’s goal is to monitor wildlife birds in the Gulf of Amvrakikos and build an intelligent system that utilizes images derived from drones equipped with thermal imagery. This system leverages state-of-the-art computer vision techniques with spatial analysis characteristics to analyze wildlife bird behavior through optical cameras and extract features for each bird. Coupled with data derived from thermal imagery as a crucial factor, it can perform a comprehensive risk assessment by correlating thermal patterns with behavioral insights, utilizing the Analytic Hierarchy Process (AHP) method. This study pioneers a novel methodology for avian health assessment using drones and identifies abnormal patterns to detect the presence of avian influenza.