Restoration of Vision of an Image in the Water

Nallagonda Amulya

Department of Computer Science and Engineering Hyderabad Institute of Technology

and Management Hyderabad,India 22e51a0581@hitam.org

Dr.S.V.Hemanth

 Department of Computer Scienceand Engineering Hyderabad Institute of Technology

and Management Hyderabad,India hemanth.sandapaga@gmail.com

Vadlakonda Rasagna

Department of Computer Science and EngineeringHyderabad Institute of Technology and Management Hyderabad,India

22e51a05C3@hitam.org

Pothagalla Rekha

Department of Computer Scienceand Engineering Hyderabad Institute of Technology

and Management Hyderabad,India 22e51a0597@hitam.org

 Malleswarapu Veera

Department of Computer Scienceand Engineering Hyderabad Institute of Technology

and Management Hyderabad,India 22e51a0570@hitam.org

Puppala Jayanth

Department of Computer Scienceand Engineering Hyderabad Institute of Technologyand Management

Hyderabad,India 22e51a05A0@hitam.org

Abstract—Underwater image quality is degraded severely because of light absorption, scattering, wavelength-dependent attenuation and other factors, and underwater images always have low contrast, color distortion and blurry phenotype. To this end, we present a Single Underwater Image Restoration framework via a Variation Model guided by Imaging Principles. In the paper a new method of contrast limited adaptive histogram equalization (CLAHE) in LAB color space is proposed to enhance the Luminance contrast and to regain real color of underwater image. To the best of our knowledge, this is the first time a webapp running GUI implemented in Python via Tkinter for single-image and multi-image restoration has been demonstrated to work across platforms. Both original and enhanced images can be viewed instantly and interaction with end users is very intuitive.The detailed and exhaustive experimental results demonstrate that the proposed method is capable of enhancing visual quality and detail visibility of underwater images in different underwater image datasets as well as restoring natural color distribution. The proposed solution generalizes existing CLAHE + variation frameworks by proposing an adaptive redchannel compensation and an interactive, cross-platform GUI for real-time single and batch underwater image enhancement.