Effective Enhancement Of Underwater Images Using Multiscale Fusion Technique

Abstract

Underwater image enhancement is an indispensable area of marine engineering research work. Underwater images typically suffer from degradation problems like low contrast, color variations, blurring, non-uniform lighting, poor visibility conditions, etc. Image enhancement is an important preprocessing step in many computers vision and graphics applications. Generally, underwater images acquired by the camera will be noisy and hazy because of refraction, absorption, and scattering effects of light. Wavelength contingent attenuation of light can be observed in deep water. Thus, the color spectrum of light with a larger wavelength attenuates more when compared to other colors spectrum of light. The major purpose of the image enhancement mechanism is to process the degraded images and results in improving the information content and tweaks visual impact. This paper is based on a multi-scale fusion-based framework with weight maps is manifested to upgrade the quality of shaded and hazed underwater imagery.