基于动态范围扩展的海底视频图像增强算法

Undersea video image enhancement algorithm based on dynamic range expansion

水下视频成像是海洋探测的关键技术,但是海底的特殊环境造成海底图像存在细节模糊、光照不均等问题。为此提出一种基于动态范围扩展的海底视频图像增强算法,该算法可以同时处理因光照不均而引起的图像曝光过度和曝光不足现象。首先,构造映射函数判断图像曝光过度和曝光不足的区域;然后,计算图像的动态范围,利用图像的"闲置空间"对图像曝光过度和曝光不足区域同时进行动态扩展;最后将图像转换到YCb Cr空间对扩展后的图像进行亮度通道的直方图均衡化。实验结果表明,和已有算法比较,新算法既可以避免传统的直方图均衡化产生的图像过增强,提高了图像的清晰度,又可以修复图像因光照不均丢失的像素信息,显著改善图像细节。

Underwater video imaging is the key technology of ocean exploration, but the special circumstance of the seabed causes to some problems on image such as fuzzy details, and uneven illumination and other issues. An undersea video image enhancement algorithm based on dynamic range expansion was proposed in this paper, which can handle image over/under-exposed phenomenon caused by uneven illumination. Firstly, The algorithm constructed a mapping function to determine the over/under-exposed region of the input image. Then, calculated the dynamic range of the image and used the“idle space”to dynamic expanding for over/under-exposed region simultaneously. Finally, The processed image was converted to YCbCr space to separate luminance channel for histogram equalization. The experiment results illustrate that the new method not only avoids unnatural results of image caused by traditional histogram equalization, but repairs missing pixel information, improves the clarity and details of image significantly.