Dataset: Mask R-CNN and U-Net model and output of coral reef halo measurements based on global multispectral satellite imagery

Final no updates expectedDOI: 10.26008/1912/bco-dmo.943698.1Version 1 (2024-12-18)Dataset Type:model results

Principal Investigator, Contact: Elizabeth Madin (University of Hawaiʻi at Mānoa)

Scientist: Simone Franceschini (University of Hawaiʻi at Mānoa)

BCO-DMO Data Manager: Karen Soenen (Woods Hole Oceanographic Institution)


Project: CAREER: Decoding seascape-scale vegetation patterns on coral reefs to understand ecosystem health: Integrating research and education from organismal to planetary scales (Coral Reef Halos)


Abstract

Reef halos are rings of bare sand that surround coral reef patches. Halo formation is likely to be the indirectly result of interactions between relatively healthy predator and herbivore populations. To reduce the risk of predation, herbivores preferentially graze close to the safety of the reef, potentially affecting the presence and size of the halo. Reef halos are readily visible in remotely sensed imagery, and monitoring their presence and changes in size may therefore offer clues as to how ...

Show more

The Mask R-CNN model was trained using the training data set, thereby automatizing the identification of reef halos from the test set of satellite images and extracting the shape of the reef halos from the imagery background (i.e., extracting both the patch reef and its surrounding halo). The Mask RCNN was trained with 3322 reef halos from 13 AOIs, while the remaining 805 halos from seven AOIs were kept to evaluate the model’s performance.

After object extraction, we automated halo measurement using a U-Net pixel classification model which discriminated the halo (sand ring) from the interior reef patch. A total of 6428 annotations from the training set were used for model training. Annotations consisted of pixel areas manually classified by users, divided into “patch reef” and “halo” classes, respectively.

The total of 6428 annotations used for the classification model was derived from splitting some of the 4127 original annotations (see related dataset). This process was implemented as part of the model training to improve pixel classification by creating diverse and multiple samples.These shapefiles were created as part of the training process and can be recreated from the original annotations already provided by processing them within the ArcGIS using the .dlpk model (as explained also in the github tensorflow).


Related Datasets

IsDerivedFrom

Dataset: Manually annotated reef halos from 6 study areas
Relationship Description: Manually annotated reef halos as input data (training & test) for Mask R-CNN and U-Net model.
Madin, E., Franceschini, S. (2024) Manually annotated reef halos based on sattelite imagery from 6 study areas as training and test data for a deep learning model. Biological and Chemical Oceanography Data Management Office (BCO-DMO). (Version 1) Version Date 2024-12-18 doi:10.26008/1912/bco-dmo.932211.1

Related Publications

Software

TensorFlow Developers. (2021). <i>TensorFlow</i> (Version v2.5.1) [Computer software]. Zenodo. https://doi.org/10.5281/ZENODO.5177380