Monitoring of Coral Reefs Using Artificial Intelligence: A Feasible and Cost-Effective Approach

Abstract

Ecosystem monitoring is central to effective management, where rapid reporting is essential to provide timely advice. While digital imagery has greatly improved the speed of underwater data collection for monitoring benthic communities, image analysis remains a bottleneck in reporting observations. In recent years, a rapid evolution of artificial intelligence in image recognition has been evident in its broad applications in modern society, offering new opportunities for increasing the capabilities of coral reef monitoring. Here, we evaluated the performance of Deep Learning Convolutional Neural Networks for automated image analysis, using a global coral reef monitoring dataset. The study demonstrates the advantages of automated image analysis for coral reef monitoring in terms of error and repeatability of benthic abundance estimations, as well as cost and benefit. We found unbiased and high agreement between expert and automated observations (97%). Repeated surveys and comparisons against existing monitoring programs also show that automated estimation of benthic composition is equally robust in detecting change and ensuring the continuity of existing monitoring data. Using this automated approach, data analysis and reporting can be accelerated by at least 200x and at a fraction of the cost (1%). Combining commonly used underwater imagery in monitoring with automated image annotation can dramatically improve how we measure and monitor coral reefs worldwide, particularly in terms of allocating limited resources, rapid reporting and data integration within and across management areas.

Coral Reef Community Changes in Karimunjawa National Park, Indonesia: Assessing the Efficacy of Management in the Face of Local and Global Stressors

“…Images were analysed using a deep learning framework to extract important biological information about the composition of reef [46,47]. In 2014, 23,135 raw images were initially collected from the area, with a further 19,085 in 2018.…”

“…A subset of 752 randomly sampled images were manually annotated at 100-point locations per image using CoralNet (http://coralnet.ucsd.edu/), where the substrate beneath each point was identified and assigned to one of 53 benthic substrate categories (see Appendix A, Table A2 for category labels and descriptions). Categories were derived from existing benthic classifications [48,49], modified to capture broad functional role but also limited by how reliably features could be identified from images by human annotators [47]. These manual annotations were used as training datasets to train a deep learning algorithm for automated image classification.…”

“…These manual annotations were used as training datasets to train a deep learning algorithm for automated image classification. The deep learning framework was then used to automatically analyse each image based on the training dataset of 752 human-annotated images [47].…”

Forecasting intensifying disturbance effects on coral reefs

“…Understanding community responses to these disturbances and associated underlying relationships requires huge amounts of data in order to estimate all the potential responses to different disturbance exposures and their interactions. Emerging technologies for environmental conservation including artificial intelligence (González‐Rivero et al, 2020), coupled with modern quantitative frameworks for collecting and analysing ecological data, can help to learn from current environmental impacts, refine our knowledge and adapt our management practices as environmental regimes change into the future. Our results here demonstrate the necessity to consider both ecological and environmental interactions when studying coral reef responses to disturbances given that the decline in HC is expected to generate different reef dynamics.…”

“…Here we used data extracted from 145,000 observations of benthic community composition collected by the XL Catlin Seaview surveys on the northern sections of the GBR between 2012 and 2017 (González‐Rivero et al, 2019, 2020). A total of 23 coral reefs were surveyed in 2012 and some of them were resurveyed in order to monitor fine‐scale changes in community composition resulting from repeated exposure to cyclones and heat stress events.…”