Coral-CRCA: A Color-Reference Chart Automation algorithm for coral bleaching visualization and severity assessment.

Area of Science:

• Marine Biology
• Ecosystem Monitoring
• Artificial Intelligence

Background:

• Coral reefs are vital marine ecosystems threatened by rising sea temperatures and pollution, leading to widespread bleaching.
• Current monitoring methods are labor-intensive and struggle with underwater image quality.
• Existing AI approaches lack fine-grained localization and full automation.

Purpose of the Study:

• To develop a fully automated algorithm for coral bleaching evaluation using underwater images and color-reference charts.
• To improve the accuracy and efficiency of coral reef health monitoring.
• To address limitations of manual annotation and image noise in current methods.

Main Methods:

• Proposed Coral Color-Reference Chart Automation (Coral-CRCA), a multi-stage algorithm.
• Implemented image denoising for underwater distortions.
• Automated coral segmentation, chart analysis, and pixel-level bleaching assessment using color similarity.

Main Results:

• Achieved 19.17% Mean Absolute Error in bleaching percentage estimation.
• Reached 96.12% binary classification accuracy (bleached/healthy).
• Demonstrated expert-level performance on 3400 field-collected images from the Arabian/Persian Gulf.

Conclusions:

• The Coral-CRCA algorithm successfully automates coral bleaching evaluation, matching expert performance.
• This AI-driven approach enhances the robustness and accuracy of monitoring degraded coral reefs.
• The developed system offers a scalable solution for assessing coral health globally.