A Novel Dehazing Algorithm Based on Median Guide Filter for Traffic Video

基于新的中值引导滤波的交通视频去雾算法

¹ School of Electronic and Control Engineering, Chang’an University, Xi’an 710064, China
² Shaanxi Road Traffic Intelligent Detection and Equipment Engineering Technology Research Center, Xi’an 710064, China
³ The 365 th Institute of Northwestern Polytechnical University, Xi’an 710065, China

Received: 2 April 2017

Abstract

Aiming at the problem that the edge preserving effect of the existing improved bilateral filtering algorithm in the process of dealing with the haze weather traffic images is not ideal, and the algorithm is time consuming and computationally inefficient, which is unacceptable in the traffic video defogging. Thus, this paper proposed a new algorithm based on median guide filter for this target. In the full paper, we explain in some detail about median guide filter; in this abstract, we just add some pertinent remarks to listing the title of each section of this paper. The six sections of this paper are:introduction(section 0), traffic image degradation model in fog and haze weather(section 1), guide filter(section 2), the design of median guide filter(section 3), Optimization of transmittance estimation(section 4), experimental result(section 5), conclusion(section 6). In section 2, we introduce the basic principle of guide filter. In section 3, we propose median guide filter and the formulas of algorithm are deduced and point out the deficiency of the traditional algorithm. In section 4, we introduce a simple method to optimize transmittance estimation. In section 5, In order to verify the effect of defog, we perform a simulation on MATLAB, the analyze of simulation results, indicates that our median guide filter algorithm can not only guarantee the effectiveness of defog, but also ensure the efficiency of calculation.

摘要

针对现有改进双边滤波算法处理雾霾天气下交通图像过程中边缘保持效果不理想且容易出现噪声斑块的问题，提出了一种基于新的中值引导滤波的去雾算法。首先，在快速引导滤波基础上改进，利用引导图像在滤波窗口的中值重新定义以像素k为中心的窗口系数 a_k 和 b_k，并应用在对滤波框内临近像素点上，避免了对引导图像和滤波图像求方差和协方差，进一步提升计算速度。去雾阶段则将含雾图像转化到 RGB 颜色空间，获取暗通道值、大气光值以及粗透射率，利用中值引导滤波算法对粗透射率进行优化，得到透射率的精确值，根据模型方程反向求解获得去雾图像。实验结果证明，与传统方法相比，新算法去雾效果强化了对边缘的检测和保持能力，更好保持了图像细节。同时，运算效率较同类算法提升较大，实时性较高，可用于交通视频处理。