Back to search

MILJØFORSK-Miljøforskning for en grønn samfunnsomstilling

A scalable WILDlife monitoring system - integrating camera sampling and artificial INTELligence with Essential Biodiversity Variables

Alternative title: Et skalerbart system for overvåking av VILLdyr - integrasjon av viltkameraer og kunstig INTELligens med essensielle biodiversitetsvariabler.

Awarded: NOK 2.4 mill.

Accurate information on biodiversity is needed to address environmental problems. Technologies like remote cameras, image classification, citizen science, and machine learning can help monitor wildlife cost-effectively, but there are challenges like high costs for manual image review. The goal of the project is to create a flexible monitoring system that uses standardized procedures from image capture to data processing. We will then test this system in different European regions using camera trapping, citizen science, deep learning, and modeling to obtain accurate biodiversity estimates. The project aims to provide a useful tool for scientists and policymakers to make informed decisions about biodiversity conservation in Europe. The Norwegian partner in the project is responsible for harmonizing image collection among study areas, designing the study, coordinating fieldwork, developing protocols, and creating a web application for data collection and processing.

Tackling the ongoing biodiversity and environmental crisis requires accurate, up-to-date information about biodiversity status, dynamics, and trends. Advances in sensing biodiversity with remote cameras and image classification technologies, citizen-science, and machine learning offer new opportunities for cost-effective wildlife monitoring. However, bottlenecks still exist for a generalized implementation of these technologies, i.e., the high costs of the manual review of images or lack of automated workflows to obtain accurate spatiotemporal data biodiversity indicators. These limitations have constrained our capacity to combine a disparity of efforts into a coherent monitoring framework and thus negatively affect conservation efforts. We will develop a scalable monitoring framework that builds on harmonized and reproducible procedures across the data cycle, from the capture of images to data processing, and estimation of Essential Biodiversity Variables. Our goal is to support the implementation of an automated monitoring system for biodiversity that integrates efficient sampling techniques and data technologies into a single framework. We will develop the system in four study areas representative of different European biogeographical regions. We will combine camera trapping, citizen science, deep learning, and hierarchical modeling to obtain unbiased estimates of Essential Biodiversity Variables. This project will contribute to European scientific excellence and capacity building by providing a biodiversity monitoring tool to scientists, managers, and policy makers to generate crucial knowledge and adopt efficient and timely management strategies. The Norwegian project partner is responsible for WP1: Harmonization of image collation among study areas, which involves development of the study design, coordination of field work, development of field protocols, and development of web application for data collection and processing.

Funding scheme:

MILJØFORSK-Miljøforskning for en grønn samfunnsomstilling