基于透射率限制修正的水下偏振图像恢复

打开文本图片集

Underwater polarization image restoration based on transmittance-limiting correction

DUAN Jin*，DONG Kehan，WANG Jialin，NI Xinyu，FANG Ruisen，ZHAO Haili （School of Telecommunications， Changchun University of Science and Technology， Changchun130022，China） * Corresponding author，E-mail： duanjin@vip. sina. com

Abstract：To address errors in transmittance estimation，an underwater polarization image restoration method based on transmittance-limiting correction is proposed. Initially，a coarse correction of the transmittance parameters is applied to suppress noise，thereby reducing interference in target imaging. Subsequently，a fine adjustment of the transmittance parameters is performed to enhance contrast and improve estimation accuracy. The corrected transmitance is ten integrated into the underwater transmission model to produce the restored image. An independently developed underwater active illumination imaging experimental device was employed to conduct restoration experiments on various targets in waters of differing turbidity levels. Experimental results demonstrate that the proposed method improves information entropy and contrast indices by averages of 33.89% and 69.01% ，respectively，compared to the original im ages.Additionaly，it outperforms other algorithms with respect to average gradient，signal-to-noise ratio， and information entropy. Target details were significantly recovered in turbid waters of varying concentrations.Under active ilumination conditions，efective restoration of underwater polarization images necessitates accurate acquisition of the backscattered degree of polarization and the light intensity at infinity. Rough estimations of these parameters may lead to transmittnce errors，resulting in considerable noise and pixel saturation in the restored images.

Key words： polarization imaging；active ilumination；underwater target image restoration；transmitance estimation

1引言

水下主动成像技术被广泛地应用于水下搜救、海洋军事目标监测、海洋资源探测等领域[1-2]。（剩余20641字）