An Image Enhancement Method Based on MSR and HDRC for Underwater Images

  • Kaiyan Zhu, Fanwei Su, Fei Wang

Abstract

An image enhancement method based on multi-scale Retinex (MSR) and high dynamic range compression(HDRC) is proposed to solve underwater image quality problem, such as low illumination and uneven illumination. Firstly, for the Halo Effect caused by MSR algorithm, a HDRC method is proposed to compress the RGB values of each channel in different degrees, which can not only suppress the Halo Effect, but also improve the color deviation effect caused by the underwater environment. Then, the low-frequency noise is filtered by Gauss high-pass filter to enhance image contrast and clarity. Finally, the chromaticity is restored by the color saturation adjustment function. Experiments are carried out onthree kinds of uneven illumination underwater images, and provethe competitive performance of the proposed method. Twotraditional underwater image evaluate metrics, underwater color image quality (UCIQE) and underwater image quality (UIQM) are selected to assess the proposed method. The experiments results showed that the proposed algorithm provides more details enhancement and more natural colorfulness restoration as compared to other existing image enhancement methods. The proposed algorithm can effectively reduceHalo Effect and noise, improve the image quality for uneven illumination underwater image.

Published
2020-12-30
How to Cite
Kaiyan Zhu, Fanwei Su, Fei Wang. (2020). An Image Enhancement Method Based on MSR and HDRC for Underwater Images. Design Engineering, 850 - 859. Retrieved from http://thedesignengineering.com/index.php/DE/article/view/1043
Section
Articles