Coral Reef Health Monitoring AI

Overview

Developed an automated computer vision system to monitor marine coral reef ecosystems over temporal imagery. The tool aligns underwater photographs taken from varying ROV angles and lighting conditions to accurately detect and quantify coral bleaching patterns and tissue decay.

Technical Methodology

• Scale-Invariant Feature Transform (SIFT): Identifies robust local keypoints invariant to illumination changes, scale variations, and camera rotation.
• Perspective Homography Alignment: Computes geometric transformation matrices to warp historical photos into exact alignment with current ROV imagery.
• Color Space Segmentation: Analyzes chromatic shifts in HSV and Lab color spaces to differentiate healthy pigmented coral tissue from bleached structures.

Source Code & Verification