目的 水下图像是海洋信息的重要载体，然而与自然环境下的图像相比，其成像原理更复杂、对比度低、可视性差。为保证不同类型水下图像的增强效果，本文提出在两种颜色模型下自适应直方图拉伸的水下图像增强方法。方法 首先，进行基于Gray-World理论对蓝、绿色通道进行颜色均衡化预处理。然后，根据红绿蓝（R-G-B）通道的分布特性和不同颜色光线在水下传播时的选择性衰减，提出基于参数动态优化的R-G-B颜色模型自适应直方图拉伸，并采用引导滤波器降噪。接下来，在CIE-Lab颜色模型，对‘L’亮度和‘a’‘b’色彩分量分别进行线性和曲线自适应直方图拉伸优化。最终，增强的水下图像呈现出高对比度、均衡的饱和度和亮度。结果 选取不同类型的水下图像作为数据集，将本文方法与融合颜色模型（ICM）、非监督颜色纠正模型（UCM）、基于暗通道先验性（DCP）的水下图像复原和基于水下暗通道先验（UDCP）的图像复原方法相比较，增强后的图像具有高对比度和饱和度。定性和定量分析实验结果说明本文提出的方法能够获得更好视觉效果，增强后的图像拥有更高信息熵和较低噪声。结论 在RGB颜色模型中，通过合理地考虑水下图像的分布特性和水下图像退化物理模型提出自适应直方图拉伸方法；在CIE-Lab颜色模型中，引入拉伸函数和指数型曲线函数重分布色彩和亮度两个分量，本方法计算复杂度低，适用于不同复杂环境下的水下图像增强。
Objective Underwater image is an important carrier of ocean information, and clear images obtained from underwater world are playing a critical role in ocean engineering such as underwater device inspection and marine biological recognition. However, compared with images captured in terrestrial environment, underwater images often present color shift, low contrast and poor visibility, because the light is absorbed, scattered and reflected by the water medium when travelling from an object to a camera in the complicated underwater environment. Existing methods cannot be effective and suitable for different types of underwater images. In order to address these problems, we propose a simple method of underwater image enhancement using adaptive histogram stretching in different color models, which can improve the contrast and brightness of the underwater image, reduce the introduction of noise and generate a relatively natural image. Method Since images are rarely color balanced in the underwater situation, we firstly pre-process the underwater image with color equalization in RGB color model based on the Gray-World (GW) assumption theory. The color equalization is only employed on the green (G) and blue (B) channels of input image to avoid an inappropriate compensation for the red (R) channel in the water which is often happened by simple color balancing. Then, we analyze the distribution characteristics of red, green and blue (R-G-B) channels which are focused on the regular range. Meanwhile, we find the rule of the selective attenuation in three channels of the underwater image that the red color is seriously affected and the wavelength of red color is the longest one, which leads to most of underwater images appearing blue-green tone. Based on the analysis and discovery, we propose an adaptive histogram stretching approach in RGB color model to adapt the different underwater images. Because the underwater images are disturbed by various factors, to reduce the impact of some extreme pixels in the process of adaptive histogram stretching, the stretching range is limited to the range [0.5%, 99.5%], and is then obtained based on the inherent characteristics similar to the variation of Rayleigh distribution. The desired range of each channel is acquired according to Rayleigh distribution theory, the image formation model and residual energy ratios of different color channels under the water. These dynamic stretching ranges have considered the characteristics of histogram distribution in hazed image and in the expected output image, simultaneously. Finally, four possible situations of the histogram stretching on the basis of the desired range are introduced to preserve the enhanced underwater images from over-stretching or under-stretching. Although the smart method built on adaptive histogram stretching will not bring obvious noise into the output image, in order to improve the contrast of the image and capture worthy details of the image, the guided filter is employed to eliminate the effect of noise. Next, in CIE-Lab color model, the ‘L’ luminance component, equivalent to the image luminance, is applied with the linear normalization in the stretched range [0.1%, 99.9%], and the brightness of the entire image receives a significant improvement. The ‘a’ and ‘b’ color components are modified to acquire color correction properly using exponential-model curve function. In the end, a color-equalized, contrast-enhanced and brightness-corrected underwater image can be produced as the perceivable output image. Result Our proposed method is evaluated by comparing with two effective non-physical methods and two state-of-the-art physical methods, qualitatively and quantitatively. The integrated color model (ICM) and the unsupervised color correction model (UCM) as typical non-physical methods are most similar to the proposed method in terms of histogram modification. The blind global histogram stretching usually tends to produce output image which contains under- or over-enhanced and under- or over-saturated areas and high noise. The DCP-based (Dark Channel Prior-based) underwater image restoration and the UDCP-based (Underwater Dark Channel Prior-based) underwater image restoration are imposed to estimate the background light (BL) and the transmission map (TM) to restore the underwater images based on optical physical model. Physical methods are just appropriate for certain underwater images enhancement and restoration under the specific circumstances and time-consuming for estimating the transmission map. Experimental results on different types of underwater images such as brown coral, underwater fishes and stones with different color tones, show that our proposed method can achieve better enhanced quality. Our method gains the highest average subjective quality score among the methods of underwater image enhancement and restoration, and further proofs that our method obtains the best visual effects. The proposed method is not only simple and effective, but also improves the contrast, details and color of the input images. In quantitative assessment, the maximum value of UCIQE means the balance of the chroma, saturation and contrast of the enhanced image in all methods; the highest value of ENTROPY represents that our method preserves the richest information and details; the lowest value of Q-MOS indicates better perceptual quality and the lowest value of MSE and the highest value of PSNR can reduce the introduction of noise when the original image is enhanced based the adaptive histogram stretching. In summary, the final results have shown that our method can recover natural underwater images, enhance the visibility of hazed images and produce high-quality underwater images. Conclusion The proposed method consists two parts: color correction and contrast enhancement in RGB color model and modification of the brightness and hue in CIE-Lab color model. In RGB color model, adaptive histogram stretching is proposed with a reasonable consideration of the distribution characters of the underwater image and the physical model of underwater image degradation. In CIE-Lab color model, stretching function and S-model curve function are adopted to modify the luminance and colors. The method proposed in this paper can be of low-complexity, and appropriate for different underwater images under complicated scenarios and more effective to enhance the visibility according to best perceptual quality, high contrast, the most information and details. Our method rewards the impressive results in the applicability and robustness compared to representative underwater image enhancement and restoration methods. In spite of the satisfactory performance, there still exist some needed improvements in our method: 1) the distance from the object to water surface and the artificial light are all ignored to influence the result of the restoration and enhancement to some extent; 2) the noise on account of the histogram stretching cannot be entirely removed based on the guided filter; 3) in deep ocean, the radiation of the natural light spreading from water surface to the object has faded away and the artificial light becomes the main source of underwater imaging. These limitations will be studied and the proposed method will be refined in the future work.