基于改进暗通道先验的海上低照度图像增强算法

An algorithm for enhancing low light images at sea based on improved dark channel priors

摘要:

针对基于暗通道先验的低照度图像增强算法在处理极端海上低光环境下图像时会存在光晕效应、色彩失真的问题，本文提出了一种基于暗通道先验的自适应海上低照度图像增强算法：首先通过选取图像类型划分指标将数据集中的图像分类，并通过Otsu方法和图像直方图分布获取图像的区域划分阈值将图像进行划分得到局部区域图，分析各类图像的局部区域图之间的关系，最后通过对不同的局部区域图采用不同的改进暗通道先验算法进行处理，将一个图像中的两个增强后局部区域图合并，最终得到整张图像的增强结果，并对增强后图像进行主客观的图像质量评价。实验结果表明，该算法解决了现有算法在处理极端海上低照度图像时存在光晕效应和色彩失真的问题，使得不同环境下的海上低照度图像都能达到较好的恢复效果。

Abstract:

Aiming at the problem of halo effect and color distortion in the dark channel a priori based low-light image enhancement algorithm when dealing with images in extreme low-light environment at sea, this paper proposes an adaptive dark channel a priori based low-light image enhancement algorithm at sea: firstly, the images in the dataset are categorized by the selection of image type classification indexes, and then the regions of the images are obtained by the Otsu method and the distribution of the histograms of the images. The local area maps are obtained by dividing the images by the Otsu method and the image histogram distribution, analyzing the relationship between the local area maps of various types of images, and finally processing different local area maps with different improved dark channel a priori algorithms, combining two enhanced local area maps of an image, and ultimately obtaining the enhancement results of the whole image, and evaluating the enhanced image subjectively and objectively in terms of image quality. The experimental results show that the algorithm solves the problems of halo effect and color distortion of the existing algorithms in processing extreme maritime low-light images, so that the maritime low-light images in different environments can achieve better recovery results.