Abstract
Abstract The rapid changes in the climate of Antarctica are likely to pose challenges to living communities, which makes monitoring of Antarctic fauna an urgent necessity. Benthos is particularly difficult to monitor, and is sensitive to local environmental changes. At the same time, long‐term monitoring is complicated by logistical factors. It is therefore urgent to develop advanced instruments to set up autonomous and long‐term monitoring programmes to obtain the lacking biological knowledge needed to understand this complex and remote marine environment. We present a pilot study to set up a non‐invasive and sustainable autonomous monitoring activity in Antarctica, leveraging on a specifically designed automated camera recording, computer vision and machine learning image processing techniques. We also present and analyse the high‐resolution image dataset acquired for an extended period of time encompassing both the summer and the Antarctic night and the corresponding transition periods. The results of this study demonstrate both the effectiveness of such an autonomous imaging devices for acquiring relevant long‐term visual data and the effectiveness of the proposed image analysis algorithms for extracting relevant scientific knowledge from such data. The presented results show how the extracted knowledge discloses dynamics of the observed ecosystems that can be obtained only through continuous observations extended in time, not achievable with the state‐of‐the‐art monitoring approaches commonly implemented in Antarctica. The success of this pilot study is a step towards the collection of continuous data near shore in Antarctic areas and in general in all the remote and extreme underwater habitats. Moreover, the presented stand‐alone and autonomous imaging device can be used for increasing the number of the monitoring sites in remote environments and when complemented with the acquisition of physical and bio‐chemical variables it can be used for obtaining data collections of great scientific value difficult to acquire otherwise.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have