We propose a multi-scale underwater image enhancement method with an optimized homomorphic filter and a RGB color correction algorithm. The proposed method includes three steps. First, an optimized homomorphic filter was developed to reduce the interference of underwater impurities on the image. The interference of impurity particles on the image is reduced, and a preliminarily optimized image can be obtained. Then, a multi-scale image enhancement algorithm is designed to enhance the color of underwater images. Finally, considering the color bias of the image, an RGB correction algorithm is proposed to correct the color of the image and further improve the image quality. Using public and simulated datasets, extensive qualitative and quantitative experiments are implemented to demonstrate the enhanced results of underwater images under different turbidity and the competitive performance of the proposed method.

我们提出了一种具有优化同态滤波器和 RGB 颜色校正算法的多尺度水下图像增强方法。所提出的方法包括三个步骤。首先，开发了一种优化的同态滤波器，以减少水下杂质对图像的干扰。减少了杂质粒子对图像的干扰，可以得到初步优化的图像。然后，设计了一种多尺度图像增强算法来增强水下图像的色彩。最后，考虑图像的颜色偏差，提出一种RGB校正算法，对图像的颜色进行校正，进一步提高图像质量。使用公共和模拟数据集，