Integrated Model for Underwater Image Enhancement and Coral Health Classification

Underwater imaging faces obstacles like limited visibility, color alteration, and image quality deterioration due to environmental factors like turbidity and inadequate lighting. These issues have an impact on oceanography, archaeology, security, and marine biology study and observation. Prevention of coral bleaching and degradation is vital for reducing coral mortality, as climate change and increased water pollution put coral reefs in threat. As a result, precise and effective methods for enhancing underwater images and identifying coral health must be created. This project aims to assess three techniques to identify the most accurate underwater image enhancing method: Underwater Image Enhancement Using White Balance, USM, and CLHE; Natural-based Underwater Image Color Enhancement through Fusion of a Swarm-Intelligence Algorithm; and Underwater Image Restoration and Enhancement Based on a Fusion Algorithm with Color Correction, Contrast Optimization, and Histogram Stretching. Peak Signal to Noise Ratio (PSNR), Structural Similarity Index Measure (SSIM), and Underwater Image Quality Measure (UIQM) are utilized to compare these methods. The selected method was used to enhance coral images for a model detecting coral health as healthy, bleached, or dead. Following the evaluation of all three approaches, NUCE obtained the highest UIQM score, while having lower PSNR and SSIM scores than the other methods. Consequently, this method was included in the YOLOv8m model for coral health classification, and it showed the highest confidence score to assess the coral's health state.