Underwater image enhancement with an adaptive dehazing framework

Chunmei Qing,Wenyou Huang,Siqi Zhu,Xiangmin Xu

DOI: https://doi.org/10.1109/ICDSP.2015.7251888

2015-01-01

Abstract:In this paper, we propose a novel method to restore and enhance the underwater degraded images. Underwater images suffer from two major problems of distortion: scattering and color change. Scattering is caused by large suspended particles and color distortion is caused by the varying degrees of attenuation encountered by light traveling in the water with different wavelengths, rendering some characteristics which are different from the foggy images in the atmosphere. Our key contribution is proposed an adaptive dahazing framework for underwater image enhancement, which includes two main parts: adaptive underwater brightness estimation and locally adaptive histogram equalization. The enhanced images are characterized by more accurate exposure of underwater images, especially in the dark regions. And improved contrast for the better details and edges are enhanced significantly in the images. Extensive experimental results demonstrate that our method is comparable to higher quality than the state-of-the-art methods.