NatureAI

National and nature parks are of crucial importance for the conservation of biodiversity, as they often represent the last remaining areas of natural ecosystems and provide habitats for rare and endangered species. To conserve biodiversity, parks must continuously monitor animal populations both spatially and temporally. Current methods, especially territory mapping by experts, are time-consuming, error-prone, not automated and are lacking available experts who can carry out this mapping. The NatureAI project is therefore developing an AI-based real-time monitoring system that uses bioacoustic data to monitor biodiversity in national and nature parks in an automated and site-specific manner. Several intelligent “smart recorders” use efficient and self-learning deep learning models to accurately identify different species. The main challenge is to quickly adapt these models to specific locations, which is achieved through optimization with deep learning techniques. For this, additional annotations are collected via an app through user feedback, building a sustainable data platform for biodiversity monitoring.

In the first phase, the project pursues the in-depth development of a detailed concept. Interfaces to users in national and nature parks will be established and possible concrete application examples selected. Based on this, recommendations for the necessary AI methods and hardware will be derived. In addition, the application for a second phase will be prepared and initial scientific work will be carried out to enable an effective and rapid entry into a possible second phase.