To cope with these issues, recent technological advances are increasingly explored to accelerate and improve the accuracy of monitoring (Andrew & Shephard, 2017 Edney & Wood, 2020 Terletzky & Ramsey, 2016). Many seabird species breed in large colonies that require expansive efforts to count manually, resulting in highly uncertain census estimations. One such indicator is the abundance and distribution of seabirds, which has the potential to provide a comprehensive measure for ecosystem health, including the lower levels of the food chain (Gregory et al., 2003 Parsons et al., 2008). The implementation of effective conservation strategies is particularly pressing given the accelerating rates of decline of global biodiversity (Butchart et al., 2010) and is tied to means of measuring biodiversity through indicators, in particular to identify possible threats and for taking effective conservation measures. Preservation of biodiversity is of great importance for the maintenance of healthy ecosystems and for human well-being (Cardinale et al., 2012 UN sustainable development goals 1 1 ). In sum, our results show that we can detect and classify the majority of 21 000 birds in just 4.5 h, start to finish, as opposed to about 3 weeks of tediously identifying and labelling all birds by hand. Our model obtains good accuracy for the most abundant species of royal terns (90% precision at 90% recall), but is less accurate for the rarer Caspian terns and gull species (60% precision at 68% recall, respectively 20% precision at 88% recall), which amounts to around 7% of all individuals present. As we employ a lightweight CNN architecture and incorporate prior knowledge about the spatial distribution of birds within the colonies, we were able to reduce the number of bird annotations required for CNN training to just 200 examples per class. By using UAVs and CNNs that allow localizing tens of thousands of birds automatically, we show that all three limitations can be addressed elegantly. Surveys to estimate breeding numbers have hitherto been carried out on foot, which is tedious, imprecise and causes disturbance. Our study area, the coast of West Africa, harbours significant breeding colonies of terns and gulls, which as top predators in the food web function as important bioindicators for the health of the marine ecosystem.
We address the task of automatically detecting and counting seabirds in unmanned aerial vehicle (UAV) imagery using deep convolutional neural networks (CNNs).
0 kommentar(er)
