Kategorie: News
High-resolution 3D tracking of coral reef fish
A study by the Leibniz Centre for Tropical Marine Research (ZMT) is using new methods to study coral reefs. Led by fish ecologist Dr Julian Lilkendey, an international research team used innovative AI technologies to analyse the movements of reef fish in the Red Sea with high precision.
The study, which was recently published in the scientific journal ‘Ecology and Evolution’ and also involved researchers from the Laboratory of Computer Science, Robotics and Microelectronics (LIRMM) at the University of Montpellier, France, and Auckland University of Technology (AUT) in New Zealand, combines stereo video technology with AI-assisted 3D tracking. The method provided detailed insights into the movement patterns and energy expenditure of two surgeonfish species in their natural habitat in the Red Sea.
The researchers first observed that the brown surgeonfish (Acanthurus nigrofuscus) showed a preference for foraging on algae growing on dead coral, while the yellowtail surgeonfish (Zebrasoma xanthurum) utilised a broader food spectrum and also ate algae found on rubble, coral rubble and sand.
The depth of the analysis was evident on several levels: Spatially, the researchers used calibrated stereo video systems to capture the three-dimensional movement of the fish as they foraged in the coral reef, which goes far beyond conventional two-dimensional observations, says first author Julian Lilkendey. A more in-depth analysis was ultimately achieved through the use of AI algorithms.
Targeted training of the AI model for species recognition
Initially, the pre-trained programme YOLOv5 (You Only Look Once version 5) was used, a neural network for object recognition in real time. For the study, YOLOv5 was fine-tuned with additional background images from the Red Sea in order to better recognise fish in the video recordings. The neural network then classified the recognised fish according to species.
The targeted training of this AI model for species recognition posed a particular challenge: ‘Because there were few specific training images for the two surgeonfish species and the region, we used media from the citizen science website “iNaturalist”,’ explains Lilkendey. ‘This allowed us to use a large number of publicly accessible photos.’
The scientists used a so-called DeepSORT algorithm (Simple Online and Realtime Tracking with a Deep Association Metric) for the subsequent three-dimensional data acquisition. ‘This algorithm enables robust multi-object tracking by tracking the recognised fish across successive video frames,’ says Lilkendey. ‘DeepSORT can track the movements of individual fish even if they briefly disappear from view or are obscured by other objects. By integrating the 3D information from the stereo image pairs, the algorithm generates precise three-dimensional movement patterns of the fish.’
Combined with an approach for modelling their energy consumption, new insights were gained into the ecology of the surgeonfish species: ‘The brown surgeonfish showed specialised feeding behaviour and preferred certain algae that grow on specific substrates, in contrast to the generalised feeding behaviour of the yellowtail surgeonfish,’ reports Lilkendey. ‘Despite their low biomass, both species make a significant contribution to grazing on the reef and utilise the energy they absorb from food for their locomotion with similar efficiency.’
Study results emphasise the role of surgeonfish in maintaining the ecological balance in coral reefs
The results emphasise the importance of niche partitioning and the role of surgeonfish in maintaining the ecological balance in coral reefs, Lilkendey continues. ‘Changes in the feeding behaviour and energy budget of surgeonfish can influence the growth of algae and the recruitment of coral larvae, which can affect the health and biodiversity of the entire reef ecosystem.’
Through the results of their analysis, the researchers were able to gain a more detailed insight into the functioning of marine ecosystems and lay the foundation for a better understanding of how energy is absorbed, converted and distributed within the reef.
Dr Lilkendey emphasises: ‘With high spatial and temporal resolution, we were able to analyse the three-dimensional movements of many fish in a coral reef simultaneously. Our new methodological approach allows us to delve deeper into the complexity of fish behaviour and the resulting energy flows.’
The research methodology also opens up new possibilities for the creation of ‘Energy Seascapes’ - detailed visualisations of the energy consumption of animals in marine ecosystems. Such mapping is important for developing effective health indicators and novel conservation measures for reefs.
Further information:
Video recordings of the study are available.
All videos are available here as a playlist on YouTube: