基于Contourlet变换和多尺度Rentinex的水下图像增强算法

  目的  水下图像是水下视觉感知技术领域应用的基础，由于水对光的吸收以及水中颗粒对光的散射作用，使水下图像具有对比度低、边缘模糊等特点，导致图像质量下降。为此，提出一种削弱水介质对水下图像影响的图像增强算法。  方法  首先，研究了水下成像数学模型，并且依据Lambert-Beer定律推导出了未退化的图像数学模型。其次，利用红色暗通道先验（RDCP）理论得到未退化图像模型中的红色暗通道透射率图像，并通过引导滤波算法对该图像进行滤波细化处理，获得了对比度比较高的未退化图像。然后，针对去模糊后水下图像对比度仍然不理想的情况，采用对比度受限直方图均衡化算法（CLAHE）提升水下图像对比度。  结果  基于常用图像评价标准，通过实验评价了本文算法的有效性，  结论  本文算法能够显著提升水下图像信息熵、对比度和平均梯度，更能凸显图片信息的细节特征。

Underwater image processing:state of the art of restoration and image enhancement methods

  Objectives  Underwater image plays an extremely important role in underwater resource exploration and non-contact inspection of underwater pipelines. However, due to absorption and scattering by water and particles in the water, the phenomenon of low contrast and blurred edges are existed in underwater image. An image enhancement algorithm is proposed to reduce the influence of water medium on underwater image.  Methods  Firstly, the mathematical model of underwater imaging was studied, and the non-degraded image mathematical model was deduced according to Lambert-Beer law. Secondly, the red dark channel transmittance image in the non-degraded image model was obtained based on the Red Dark Channel Prior (RDCP) theory, and the non-degraded image with high contrast was obtained by filtering and thinning the image with the guided filter algorithm. In addition, since the image contrast was still not ideal after deblurring, the Contrast Limited Adaptive Histogram Equalization (CLAHE) algorithm was employed to improve the image contrast.  Results  Based on the commonly used image evaluation criteria, the effectiveness of the proposed algorithm in this paper is evaluated by experiments.  Conclusions  It is found that the algorithm proposed in this paper can significantly improve the information entropy, contrast and average gradient of underwater images, and more information is obtained from the underwater image.