一种基于色彩恒常理论的薄雾图像增强方法

针对薄雾天气下的图像对比度较低,以及光照不均引起的颜色退化失真问题,提出了一种基于色彩恒常理论的薄雾图像增强方法。该方法将Frankle_McCann Retinex(FMR)算法与光照补偿算法系统地结合。首先利用FMR算法来提高图像对比度,再采用光照补偿方法,将彩色图像从RGB空间转换至HSV空间,对亮度分量进行直方图均衡处理;再转回RGB空间,以此提高图像亮度同时保留色彩真实性。实验表明,此方法比传统FMR算法能更有效去除色彩失真,改善退化细节,获得良好视觉效果。

A method for Images Defogging Enhancement Based on Color Constancy Theory

Under the fog weather, the color image taken by device is degraded by atmospheric. According the features, a combined method is proposed based on the color constancy theory Frankle_McCann Retinex (FMR) and illumination compensation algorithm. Firstly, the photos are transformed from the RGB space to the HSV space, Then, the method of histogram equalization is used to realize the adjustment of the luminance and then back to the RGB space to complete the FMR enhancement. The experiments show that this method is better than the traditional FMR algorithm effectively removing the color distortion caused by uneven illumination, further improving the degradation details, and getting a nice visual effect.