摘要
海面及天空区域对暗原色先验失效,通过控制参数在明亮区域求得透射率,从而改进了海上视频图像暗原色先验去雾算法,避免了图像的色彩失真。并且,根据海域旷阔的特点,推导了单目视觉的视频图像景深分布模型,结合大气散射模型,提出了基于单幅视频图像景深的图像去雾法。通过多个海上及陆上雾天场景降质视频图像的处理,其结果验证了算法的有效性。
In view of the problem that dark channel priori fails in offing and sky areas, the dark channel priori de hazing method is adapted to maritime video images by correcting the color distortion of images through finding trans mission rate in bright regions. For maritime images featuring deep and wide viewing, the depth ofield distributing model of monocular vision is established. By combining the depth-of-field model with the atmosphere scatter model, a single frame video image de-hazing method is proposed. The effectiveness of this algorithm is validated by the dehazing results of processing several degraded images, maritime or overground.
出处
《中国航海》
CSCD
北大核心
2013年第1期13-17,共5页
Navigation of China
基金
国家自然科学基金(60975016
61002052)
关键词
船舶
舰船工程
大气散射模型
去雾
景深模型
暗原色
ship, naval engineering
atmosphere scatter model
de hazing
depth-of field model
dark channel priori
