水下图像增强和复原方法研究

Research on Underwater Image Enhancement and Restoration Methods

针对水下机器人在海洋水下环境拍摄的图像普遍存在颜色失真、低对比度、场景细节模糊等问题,提出一种基于直方图均衡化的泊松方程原理方法的组合算法。首先对图像的绿色通道和蓝色通道进行泊松方程原理方法处理,达到保持图像梯度目的,然后对图像每个通道进行均衡化处理,改善图像颜色失真,再对图像进行颜色衰减补偿算法处理,减轻红色通道过补偿现象,最后再基于泊松方程原理方法对图像进行处理,从而增强图像的对比度、清晰度,并恢复出更多场景特征信息。实验证明,该组合算法能很好地改善水下图像颜色失真,解决图像对比度低、场景细节模糊等问题,处理后的图像符合人眼视觉观察效果,能很好地改善图像质量。

A combined algorithm based on histogram equalization and the Poisson equation principle method is proposed to address the issues of color distortion,low contrast,and blurred scene details commonly observed in images captured by underwater robotics in marine environments.Firstly,the algorithm applies the Poisson equation principle method to the image′s green and blue channels,conserving their gradients.Then,histogram equalization is performed on each channel to reduce color distortion.In addition,a color attenuation compensation algorithm is applied to the image to prevent red channel overcompensation.The image is subsequently reprocessed using the Poisson equation principle method to improve contrast,sharpness,and the scene′s feature information.The experimental results demonstrate that this combined algorithm effectively improves color distortion in underwater images,resolves issues with low contrast and blurred scene details,and improves low contrast and blurred scene details.The processed images have an effect that is visually pleasing and substantially improve image quality.