Application of drones and artificial intelligence in flocking birds

The emergence of unmanned aerial vehicles (UAVs) has expanded the boundaries of traditional bird survey methods, enabling researchers to reach many previously inaccessible targets. However, UAVs can potentially disrupt bird species and remain a concern for researchers. This study indicates that among 11 species of wintering waterbirds, there is a significant negative correlation between bird size and the alert distance from UAVs. In the case of breeding Bridled Terns (Thalasseus bernsteini), their alert distance to UAVs during chick stage is significantly greater than during egg incubation stage. Additionally, experiments on Eurasian Teals (Anas crecca) and Black-winged Stilts (Himantopus himantopus) have shown a trend of shortening their alert distances to UAVs with repeated testing over a short period.Population estimation of a large breeding colony is an important indicator for planning conservation efforts. However, traditional methods of estimating their population is both time-consuming and labor-intensive, limiting the research team’s energy for other studies. To address this situation, this study has developed an artificial intelligence model capable of automatically counting the population of terns. By replacing traditional counting methods with this model, work efficiency can be improved, allowing the research team to contribute to a broader range of research endeavors.