改进型Retinex算法的水利水电图像特征提取与分析

Extraction and analysis of water conservancy and hydropower image features based on improved Retinex algorithm

针对我国水利水电工程中水下探测过程采集到的图像质量不佳、画面昏暗的问题,文中使用基于导向滤波的Retinex图像增强算法,对采集到的水下图像进行复原处理,将水下图像RGB色彩空间转换为HSV色彩空间,采用直方图均衡化和颜色校正技术,解决了水下图像噪声较大和边缘模糊的问题。使用同态滤波算法对图像的高频信息和低频信息进行处理,抑制图像光照较强的区域。实验结果显示,该研究的算法在优化图像增强算法方面具有突出的技术优势。

Based on the problem of poor image quality and dim images collected during underwater detection in our country’s water conservancy and hydropower projects,this paper uses the Retinex image enhancement algorithm based on guided filtering to restore the collected underwater images and convert the underwater images color space from RGB to HSV.Besides,the histogram equalization and color correction technology are used to solve the problem of large noise and blurred edges in underwater images.The homomorphic filtering algorithm is used to process the high-frequency information and low-frequency information of the image,and suppress the areas with strong illumination of the image.The experiment results show that the algorithm in this study has outstanding technical advantages for optimizing the image enhancement algorithm.