发光目标物团雾视频数值分析

高速公路水平能见度视频图像监测技术正取得快速发展，已成为团雾天气监测、预警的重要手段之一。为有效预防公路团雾交通事故发生，参照高速公路团雾定义和预警等级的规定，基于公路团雾多发路段视频和能见度资料，统计分析了路面固定距离的发光目标物在团雾和非团雾不同能见度环境下的图像光亮度参数，探究了团雾路段不同距离处发光体亮度数值与水平能见度数值相关变化关系。进而基于团雾路段视频特征图像区域的亮度参数变化与水平能见度的相关性分析，给出利用数值分析结果测算水平能见度的可行方法，为建立高速公路团雾预警模型提供思路和实现途径。研究结果表明：(1)通过视频图像的不同区域亮度变化进行能见度监测具有科学性、创新性、可行性，可实现24 h不间断监测预警。(2)引入发光目标物光亮度和环境光亮度的对比分析，可以解决一天中不同时间段图像光亮度变化较大的问题。(3)通过对固定距离发光目标物的监测来实现公路团雾的监测，可解决团雾水平尺度难以直接仪器监测的难题。(4)发光目标物视频团雾监测方法的监测结果可与互联网、地理信息系统、导航信息系统结合，从而实现快速、准确发布预警，减少交通事故的目的。

Numerical analysis of local dense fog by video image monitoring of luminous objects

The video image monitoring technology of highway horizontal visibility is developing rapidly, and has become one of the important means of local dense fog weather monitoring and early warning. In order to prevent road fog traffic accidents effectively, referring to the definition of local dense fog and the regulation of early warning level, this paper statistically analyzes the luminance parameters of luminous objects at fixed road distance in different visibility environments of dense fog and non-dense fog based on the video and visibility data of road fog-prone sections, and explores the correlation between the luminance value of luminous objects and the horizontal visibility value at different distances in dense fog section. Then, based on the correlation analysis between the luminance parameter changes and the horizontal visibility of the video feature image, a method to estimate the horizontal visibility using the numerical analysis results is presented, which provides a new idea and realization way for the establishment of local dense fog weather early warning. The results show that: (1) It is scientific, innovative and feasible to monitor the visibility through monitoring the luminance change in different areas of the video image, which can realize 24 h continuous monitoring and early warning. (2) By introducing the contrast analysis of the luminance of the luminous object and that of the environment, the problem that the luminance of the image varies greatly at different time of the day can be overcome. (3) The monitoring of local dense fog can be realized by monitoring the luminous objects at a fixed distance, which can solve the problem that the horizontal scale of local dense fog is difficult to be directly monitored by instruments. (4) The monitoring results of the luminous target video monitoring method can be combined with the internet, geographic information system and navigation information system, so as to realize the purpose of issuing early warning quickly and accurately, and reducing traffic accidents.