基于Retinex的低照度图像增强算法研究

Research on Low Illumination Image Enhancement Algorithm Based on Retinex

针对低照度图像亮度小、细节信息不突出等特点,本文使用一种改进的Retinex算法。首先对低照度图像进行同态滤波处理,使原图整体亮度增强;然后将其变换到HSV颜色空间,对亮度分量V使用改进的MSR变换进行增强;最后图像变换到RGB空间得到增强图像。实验结果表明,本文所采用的算法能够很好地保持图像细节信息,提高原图局部对比度,达到低照度图像增强的目的。

Aimed at features including less luminance of low illumination image and not prominent details,this article employs a improved Retinex algorithm.First,low illumination image is processed by homomorphic filtering to enhance the overall luminance of original image.Then,we switch it to HSV color space,using improved MSR switching to enhance luminance component V.At last,the image is switched to RGB space to acquire an enhanced image.The outcome of experiment indicates that the algorithm used by this paper can maintain detail information of image well,increase the local contrast of original image,thereby enhancing low illumination image.