Underwater Image Enhancement Using FPGA-Based Gaussian Filters with Approximation Techniques

The major challenge in marine environment imaging lies in addressing the haziness induced by natural phenomena, such as absorption and scattering in underwater scenes. This haze significantly impacts the visual quality of underwater images, necessitating improvement. This paper presents a novel approach aimed at enhancing the efficiency of Gaussian filters for reducing Gaussian noise in underwater images. The method introduces a pipeline structure in the Gaussian filter implementation and evaluates the influence of employing approximate adders on overall performance. Simulation results reveal a notable speed enhancement exceeding 150%, coupled with a substantial reduction in power consumption exceeding 34%. However, these advantages are tempered by an increase in spatial requirements. The study recognizes the inherent tradeoff between output quality and power, highlighting the applicability of the proposed design in error-resilient applications, particularly in image and video processing domains. In essence, the presented approach offers a compelling solution where the benefits of accelerated speed and reduced power consumption outweigh spatial constraints, contributing to the advancement of underwater image enhancement techniques.