Optimization of Fusion Weighting for CLAHE-MSR Using Artificial Bee Colony for Image Enhancement in the Seagrass Bed Ecosystem

Abstrak

Seagrass ecosystems are habitats rich in biodiversity. Monitoring seagrass ecosystems using underwater imagery is hindered by image degradation due to light absorption and scattering in water. These conditions can hinder visual analysis and subsequent computer vision-based processing. Therefore, a practical, stable, and appropriate method for improving underwater image quality is needed for limited data conditions. This study proposes an underwater image enhancement strategy that fuses CLAHE and MSR with fusion weights optimized by the Artificial Bee Colony (ABC) algorithm. Weight optimization was performed using the UIEB dataset with UIQM as the fitness function. The optimal weights were then applied to seagrass images, including monospecific and mixed-vegetation images from natural coastal environments. The results of the experiment show that CLAHE–MSR fusion with optimal weighting produces more balanced image quality than the separate application of CLAHE or MSR. Evaluation using the UIQM and its decomposition indicates that the fusion strategy effectively combines local contrast enhancement, color stability, and image sharpness. Additional analysis using information entropy (IE) also indicates that the fused image retains a richer information distribution without incurring excessive noise. Overall, this study shows that the traditional optimization-based image enhancement approach remains relevant and applicable as a preprocessing stage to support seagrass ecosystem monitoring and underwater image processing using computer vision, particularly in conditions where data and computational resources are limited.

Identifier

• DOI: https://doi.org/10.62527/joiv.10.1.4887
• OpenAlex: https://openalex.org/W7128914965