水面波动和水体湍流退化图像的复原方法  

Correction methods for water fluctuation and underwater turbulence degraded imaging

作　　者：鲁啸天 杨天鸣[1] 金伟其[1] 刘敬[1] 温仁杰 

机构地区：[1]北京理工大学光电学院光电成像技术与系统教育部重点实验室,北京100081

出　　处：《应用光学》2017年第1期42-55,共14页Journal of Applied Optics

基　　金：国家自然科学基金重点项目(61231014)和面上项目(61575023);总装预研基金(9140A02060415BQ01005)

摘　　要：水面波动对水下图像造成的畸变和水体湍流造成的模糊等严重制约了空中对水下目标探测、水下透空目标警戒、海上搜救等的应用,实现畸变和湍流校正具有重要的意义。综述了复原水面波动和水体湍流引起的图像失真的研究进展,给出了基于幸运块(lucky patch)选择的校正、基于图像配准的校正、基于水面波形估计的校正和基于图像退化模型的校正四大类方法的特点及典型的图像复原效果,并分析了复原水面波动和水体湍流退化图像复原方法进一步深入研究的内容。Underwater image suffers from distortions and blurs due to water fluctuations and underwater turbulence that restricts the development of underwater surveillance, underwater target alert in the air, maritime search severely. The realization of distortion and turbulence correction has great significance. Most recent developments for the degraded image by water fluctuations and underwater turbulence are reviewed in this paper, and four methods and typical image restoration results based on lucky patch, image registration, water-waveestimation and image degradation model are summarized accordingly. Further research directions for restoring underwater degraded image are analyzed at the end of the paper.

关 键 词：水面波动 水体湍流 图像复原 图像配准 幸运块 

分 类 号：TN911.73[电子电信—通信与信息系统]