王丹, 张子玉, 赵金宝, 梁文峰, 杨谢柳, 范慧杰, 唐延东. 基于场景深度估计的自然光照水下图像增强方法[J]. 机器人, 2021, 43(3): 364-372. DOI: 10.13973/j.cnki.robot.200275
引用本文: 王丹, 张子玉, 赵金宝, 梁文峰, 杨谢柳, 范慧杰, 唐延东. 基于场景深度估计的自然光照水下图像增强方法[J]. 机器人, 2021, 43(3): 364-372. DOI: 10.13973/j.cnki.robot.200275
WANG Dan, ZHANG Ziyu, ZHAO Jinbao, LIANG Wenfeng, YANG Xieliu, FAN Huijie, TANG Yandong. An Enhancement Method for Underwater Images under Natural Illumination Based on Scene Depth Estimation[J]. ROBOT, 2021, 43(3): 364-372. DOI: 10.13973/j.cnki.robot.200275
Citation: WANG Dan, ZHANG Ziyu, ZHAO Jinbao, LIANG Wenfeng, YANG Xieliu, FAN Huijie, TANG Yandong. An Enhancement Method for Underwater Images under Natural Illumination Based on Scene Depth Estimation[J]. ROBOT, 2021, 43(3): 364-372. DOI: 10.13973/j.cnki.robot.200275

基于场景深度估计的自然光照水下图像增强方法

An Enhancement Method for Underwater Images under Natural Illumination Based on Scene Depth Estimation

  • 摘要: 针对水下图像模糊、对比度低且色彩失真的问题,结合修正的水下成像模型,提出一种基于场景深度估计的自然光照水下图像增强方法.首先,依据自然光照条件下水下图像场景亮度与场景深度总体成正比的先验理论,对图像的亮度信息进行最小值滤波和软抠图处理以实现场景深度估计;然后,结合暗通道先验知识和场景深度信息进行离散像素点的后向散射分量估计,根据修正的水下成像模型来拟合和去除后向散射;最后,采用基于色适应的颜色校正方法对直接分量进行色偏校正,利用线性拉伸方法来提升图像的亮度和对比度.水下多场景条件下的实验结果表明,本文方法可有效地去除后向散射引起的雾样模糊,提高图像对比度并校正颜色偏差.

     

    Abstract: To address the problems of blur, low contrast and color cast in underwater images, an enhancement method for underwater images under natural illumination based on scene depth estimation is proposed by combining the refined underwater imaging model. Firstly, the brightness information of an image is used to estimate the scene depth by minimum filtering and soft matting method, according to the prior theory that the brightness of the underwater scene under natural illumination conditions is generally proportional to its scene depth. Secondly, the backscatter of some discrete pixels is estimated by using the dark channel prior and the scene depth information, and then the backscatter in the whole image is fitted and removed by combining the refined underwater imaging model. On this basis, the chromatic-adaptation-based color correction method is applied to the direct component of the image to remove the color cast. In addition, the linear stretching method is adopted to improve the brightness and contrast of the restored image. The experiment results in various underwater scenes show that the proposed method can effectively remove the foggy blur caused by backscatter, and improve the contrast and correct the color cast of underwater images.

     

/

返回文章
返回