交通运输系统工程与信息

交通运输系统工程与信息 ›› 2016, Vol. 16 ›› Issue (4): 88-94.

• 智能交通系统与信息技术 • 上一篇    下一篇

雾霾天气情况下的交通标志检测

薛玉利   

  1. 山东青年政治学院信息工程学院，山东省高校信息安全与智能控制重点实验室，济南250103
  • 收稿日期:2015-12-02 修回日期:2016-03-25 出版日期:2016-08-25 发布日期:2016-08-26
  • 作者简介:薛玉利（1981-），女，山东济南人，讲师，硕士.

Traffic Sign Detection under Fog and Haze Weather

XUE Yu-li   

  1. Key Laboratory of Information Security and Intelligent Control in Universities of Shandong, School of Information Engineering, Shandong Youth University of Political Science, Jinan 250103, China
  • Received:2015-12-02 Revised:2016-03-25 Online:2016-08-25 Published:2016-08-26

PDF

被引次数

1

摘要：

针对雾霾天气不利于交通标志检测的问题，提出一种在雾霾天气下快速实现 交通标志检测的算法.首先,通过暗原色原理对雾霾天气中获得的图像进行去雾处理，得 到对比度增强的图像；然后，将图像转换为归一化红蓝图像，在不同阈值对该图像进行二 值化，提取其连通区域，如果连通区域在几个阈值下的二值化图像均能保持形状不变，则 选作感兴趣区域；最后，利用交通标志的形状信息将干扰区域去除，得到交通标志的检测 结果.实验结果表明，雾霾天气的图像经过去雾处理后对比度增强，检测效率有明显提高； 与使用单一阈值检测算法相比，最大稳定极值区域的检测效率更高.

关键词: 智能交通, 交通标志检测, 暗原色原理, 最大稳定极值区域, 雾霾天气

Abstract:

To solve the crucial issue of efficient traffic sign detection under fog and haze environment, a fast and robust detection algorithm is proposed in this paper. First, enhanced images are obtained by defogging with dark channel prior principle, which are further converted into normalized red- blue images. Then, different thresholds are applied to obtain series of binary images, where connected regions that remain unchanged are selected as the region of interest. Finally, interference regions are further removed by shape constraints of traffic signs, and the final detection results are obtained. Abundant experimental results show that image qualities under fog and haze weather are enhanced after defogging, and traffic sign detection performance is improved significantly. The detection efficiency is higher in the region of maximum stability comparing with the single threshold detection algorithm.

Key words: intelligent transportation, traffic sign detection, principle of dark channel prior, MSERs, fog and haze weather

中图分类号: