ESR 39:91-104 (2019)  -  DOI: https://doi.org/10.3354/esr00958

ABSTRACT: There is increasing demand for efficient ways to process large volumes of data from visual-based remote-technology, such as unmanned aerial vehicles (UAVs) in ecology and conservation, with machine learning methods representing a promising avenue to address varying user demands. Here, we evaluated current trends in how machine learning and UAVs are used to process imagery data for detecting animals and vegetation across habitats, placing emphasis on their utility for endangered species. We reviewed 213 publications that used UAVs at 256 study sites, of which just 89 (42%) used machine learning to assess the visual data. We evaluated geographical and temporal trends and identified how each technology is used at a global scale. We also identified the most commonly encountered machine-learning methods, including potential reasons for their limited use in ecology and possible solutions. Thirteen out of the 17 habitats defined by the International Union for Conservation of Nature (IUCN) habitat classification scheme were monitored using UAVs, while 12 habitats were monitored using both UAVs and machine learning. Our results show that, while machine learning is already being used across many habitat types, it is primarily restricted to more uniform habitats at present. Out of 173 plant and animal species monitored using UAV surveys, 30 were of conservation concern, with machine learning being used to assess UAV imagery data for 9 of these species. In conclusion, we anticipate that the joint use of UAVs and machine learning for ecological research and conservation will expand as machine learning methods become more accessible.

KEY WORDS: Meta-analysis · Computer vision · Video analysis · Big data · Classifier · Deep learning · BRUVs · ROVs

Full text in pdf format Supplement 1 Supplement 2

Cite this article as: Dujon AM, Schofield G (2019) Importance of machine learning for enhancing ecological studies using information-rich imagery. Endang Species Res 39:91-104. https://doi.org/10.3354/esr00958 Export citation Share:    Facebook - - linkedIn