Particle Swarm Optimization Framework for Underwater Image Enhancement in Big Data Context

Underwater images are characterized by low-quality, poor, and non-uniform illumination due to light absorption, scattering, and suspended particles. Enhancing these images is significant in improving the image clarity and contrast, which are essential for marine biology, underwater exploration, and environmental monitoring tasks. This research proposes a Particle Swarm Optimization (PSO) framework, integrated with classical enhancement algorithms, to address these challenges in a Big Data environment. The PSO algorithm optimizes contrast and clarity while preserving fine image details, ensuring high-quality enhancements suitable for downstream analytics. The enhancement pipeline incorporates color cast neutralization for superior color channels and dynamic adjustments of gain factors (J and K) for inferior color channels. The histogram is separated into two regions at the average point of mean and median values of the image histogram. Histogram stretching is applied over the entire dynamic range and is combined to create a new histogram of the enhanced image. Mean equalization-based PSO is applied and finally, the edges and details are enhanced by unsharp masking. Compared to traditional methods, experimental results demonstrate remarkable visual and quantitative improvements. The algorithm was tested on Heron Island Coral Reef Public Database (HICRD) and satisfactory results were obtained. The proposed PSO-based enhancement technique increases the interpretability and reliability of an underwater image, validated in terms of performance metrics.