基于改进Retinex的城市交通图像增强

为提高在雾霾天气条件下获取的城市交通图像的对比度和清晰度,本文提出了一种基于引导滤波平滑约束和分数阶积分的改进Retinex的图像增强方法.Retinex理论把一幅图像视为照射分量与反射分量乘积的形式.首先,利用引导滤波作为Retinex算法的目标函数的平滑约束项,通过最小化该目标函数获取照射分量;反射分量为原始图像与照射分量的比值,其主要体现图像的局部细节,对反射分量进行分数阶积分掩膜,在去除噪声干扰的同时保留图像的细节信息;最后,将处理后的反射分量与照射分量相乘,得到最终的增强图像.实验结果表明,本文方法在图像平滑和细节保持方面得到了很好的平衡,有效地提高了雾霾天气条件下城市交通图像的对比度和清晰度,使得增强后的图像更加真实自然.

An Urban Traffic Image Enhancement Method Based on Modified Retinex

In order to improve the contrast and clarity of the urban traffic image obtained under haze weather condition, a modified Retinex image enhancement method based on guided filtering constraint and fraction integral is proposed in this paper. According to the Retinex theory, an image can be regarded as the product of the illumination component and the reflection component. Firstly, the guided filtering is applied as the smoothness constraint of the objective function of the Retinex algorithm, and the illumination component is obtained by minimizing the objective function; then, the fractional integral mask is used to remove noise and preserve detailed information of the reflection component, which is acquired using the ratio between original image and illumination component; At last, the final enhanced image is the product of the illumination component and the processed reflection component. The experimental results show that the proposed method could keep a good balance in image smoothing and detailed information preserving, improve the contrast and clarity of the highway traffic image, and make the enhanced image more natural.