基于水下图像光学成像模型的清晰化算法

摘要

To overcome the problems of underwater images such as fuzzy texture details, low contrast and non-illumination, the underwater images imaging process was first analyzed and then a visibility enhancing algorithm was proposed-Underwater image optical imaging model was used in the low-frequency sub-band, where image with medium scattering light was estimated and eliminated using Gaussian blur, and then attenuation factor was adjusted based on local complexity method to enhance adaptively low frequency sub-image. Non-linear transform for enhancing image was used in the high-frequency sub-band, which further enhanced the high frequency information and effectively restrained the noise magnification. The experimental results show that the algorithm can effectively deal with the problem of image blurring and non-illumination, and the running time is less than that of restoration algorithm for degraded underwater images based on wavelet transform.%针对水下图像的纹理细节模糊、对比度低以及图像光照不均问题,通过分析水下图像的成像过程,提出一种水下图像清晰化算法.在小波域的低频子带上结合水下图像光学成像模型,先利用高斯模糊对介质散射光进行估计与去除,再采用基于局部复杂度的方法调整衰减因子,对衰减低频子图进行自适应增强；在高频子带上采用非线性变换的增强方法,进一步增强了高频信息并有效地抑制了噪声的放大.实验结果表明该算法对解决水下图像模糊和光照不均问题具有较好的效果,与基于小波变换的水下降质图像复原算法相比,具有较高的实时性.