Search

Reef-building corals are ecosystem engineers that compete with other benthic organisms for space and resources. Corals harvest energy through their surface by photosynthesis and heterotrophic feeding, and they divert part of this energy to defend their outer colony perimeter against competitors. Here, we hypothesized that corals with a larger space-filling surface and smaller perimeters increase energy gain while reducing the exposure to competitors. This predicted an association between these two geometric properties of corals and the competitive outcome against other benthic organisms. To test the prediction, fifty coral colonies from the Caribbean island of Curaçao were rendered using digital 3D and 2D reconstructions. The surface areas, perimeters, box-counting dimensions (as a proxy of surface and perimeter space-filling), and other geometric properties were extracted and analyzed with respect to the percentage of the perimeter losing or winning against competitors based on the coral tissue apparent growth or damage. The increase in surface space-filling dimension was the only significant single indicator of coral winning outcomes, but the combination of surface space-filling dimension with perimeter length increased the statistical prediction of coral competition outcomes. Corals with larger surface space-filling dimensions (Ds > 2) and smaller perimeters displayed more winning outcomes, confirming the initial hypothesis. We propose that the space-filling property of coral surfaces complemented with other proxies of coral competitiveness, such as life history traits, will provide a more accurate quantitative characterization of coral competition outcomes on coral reefs. This framework also applies to other organisms or ecological systems that rely on complex surfaces to obtain energy for competition.

The semantic segmentation of underwater imagery is an important step in the ecological analysis of coral habitats. To date, scientists produce fine-scale area annotations manually, an exceptionally time-consuming task that could be efficiently automatized by modern CNNs. This paper extends our previous work presented at the 3DUW’19 conference, outlining the workflow for the automated annotation of imagery from the first step of dataset preparation, to the last step of prediction reassembly. In particular, we propose an ecologically inspired strategy for an efficient dataset partition, an over-sampling methodology targeted on ortho-imagery, and a score fusion strategy. We also investigate the use of different loss functions in the optimization of a Deeplab V3+ model, to mitigate the class-imbalance problem and improve prediction accuracy on coral instance boundaries. The experimental results demonstrate the effectiveness of the ecologically inspired split in improving model performance, and quantify the advantages and limitations of the proposed over-sampling strategy. The extensive comparison of the loss functions gives numerous insights on the segmentation task; the Focal Tversky, typically used in the context of medical imaging (but not in remote sensing), results in the most convenient choice. By improving the accuracy of automated ortho image processing, the results presented here promise to meet the fundamental challenge of increasing the spatial and temporal scale of coral reef research, allowing researchers greater predictive ability to better manage coral reef resilience in the context of a changing environment.