水下图像的颜色校正及对比度增强算法研究

Research on Color Correction and Contrast Enhancement Algorithm of Underwater Image

通常水对光的吸收以及水中的悬浮物会使水下图像出现色偏、对比度差、光照不均匀等问题。为解决这些问题,首先,采用统计学的方法来对水下色偏图像进行颜色矫正;其次,通过对比度代码图像(CCI)改进的局部自适应对比度增强(ACE)算法以及限制对比度自适应直方图均衡(CLAHE)算法的处理后,得到两个不同版本的对比度增强图像;最后,为获得高质量水下增强图像,借助多尺度融合算法,将可变邻域自适应对比度细节增强图像和局部限制对比度增强及亮度均衡图像融合。在890张有可参考图像的UIEB数据集中,通过对现有的一些算法从定性和定量的角度进行比较分析,结果表明,所提出的算法能够有效地改善图像色彩并提高图像对比清晰度。

Usually,the absorption of light by water and suspended matter in water can cause color skewing,poor contrast,un⁃even illumination and other problems in underwater images.In order to solve these problems,firstly,a statistical method is used to correct the color of the underwater color cast image.Secondly,two different versions of contrast-enhanced images are obtained after processing by the improved local adaptive contrast enhancement(ACE)algorithm of the contrast code image(CCI)and the limited contrast adaptive histogram equalization enhancement(CLAHE)algorithm.Finally,in order to obtain high-quality underwater en⁃hanced images,a multi-scale fusion algorithm is used to fuse the variable neighborhood adaptive contrast detail enhanced image and the locally limited contrast enhanced and brightness equalized image.In the UIEB dataset with 890 reference images,by com⁃paring and analyzing some existing algorithms from both qualitative and quantitative perspectives,the results show that the proposed algorithm can effectively improve image color and image contrast sharpness.