基于引导滤波的Retinex快速夜间彩色图像增强技术

针对目前基于Retinex原理的图像增强技术在对夜间彩色图像处理中存在颜色失真、在高对比度边缘区域存在“光晕伪影”等问题, 提出了一种基于引导滤波的Retinex快速夜间彩色图像增强的算法.该算法首先构造优化的颜色恢复函数, 然后对夜间彩色图像在色调、饱和度、亮度(hue, saturation, value, HSV)颜色空间下的亮度图像采用具有边缘保持功能的引导滤波进行平滑处理, 估计场景的光照分量, 进而获得仅包含物体本身特性的反射分量图像, 经过Gamma校正与颜色恢复后最终实现夜间彩色图像增强.实验结果表明, 本文提出的算法增强效果显著、效率较高, 同时具有颜色保真、抑制过增强和能较好保留边缘细节信息等特性.

Rapid Retinex Algorithm for Night Color Image Enhancement Based on Guided Filtering

A rapid Retinex algorithm based on guided filtering is proposed to the question of color distortion and the apparition of halo artifacts in the presence of high contrast edges using exiting Retinex image enhancement. First, this algorithm constructs a optimization color restoration function, and processes brightness image of hue, saturation, and value (HSV) color space using the guided filter that has the function of edge-preserving to estimate the scene illumination, then it obtains the reflection image, finally realizes the night color image enhancement after Gamma correction and color restoration. Experimental results show that this algorithm has good enhancement result and high efficiency, which also has the characteristics of color restoration, enhancement control and edge preserving.