一种利用前景模型的水下图像增强算法

Underwater Image Enhancement Algorithm Using Foreground Modeling

在水下图像成像过程中,由于水体对光线的选择性吸收和光的散射作用,水下图像经常出现颜色失真以及图像模糊,传统的去雾算法和简单的色彩校正用于水下图像时效果欠佳.根据先去除图像模糊后去除颜色失真的思路,结合人们对水下图像的认知,本文提出了一种基于前景模型的水下图像增强方法.利用光在水中的衰减特性,根据各颜色通道衰减系数之间的关系修正通道增益,提出适用于水下图像的色彩校正方法.另外,改进的背景光估计方法可以有效的避免水下图像出现过曝光.主观和客观的实验结果均表明,该方法在增强图像对比度和提升清晰度方面效果良好,有效的解决了图像模糊和颜色失真的问题.

The traditional dehazing methods and simple color correction fail when used in underwater image because of the color distor- tion and image blurring, caused by the selective absorption of water and light scattering in the imaging process of underwater. According to the ideal of removing the color distortion after the image blurring removal, combing with human perception of underwater images, a new method based on foreground modeling is proposed in this paper. Due to the attenuation of light in water,a color correction approach that applied to underwater is proposed. The relationship of attenuation between each color channel is used to obtain a more accurate gain estimation. In addition, the improved estimation of background light can avoid overexposure in underwater images. The performance of the proposed method is evaluated objectively and subjectively, both experimental results show that the method can improve the contrast and visibility of underwater images effectively ,and the images after enhancement have no problem of image blurting and color distortion.