摘要
针对现有交通状态检测算法无法适应城市道路复杂交通的问题,提出一种新的基于Haar-like和时空信息的交通状态区域提取算法。该算法首先采用基于Haar-like特征的车辆检测算法、边缘检测法和帧差法分别提取路面车辆、空域纹理和时域纹理的三种信号;然后将提取到的三种信号进行统计分析,获得准确的交通状态区域。将该算法与基于车辆检测的交通状态检测算法和基于帧差法的交通状态检测算法在远距离小目标、遮挡车辆和混合交通的复杂交通场景中进行对比实验。实验结果表明,该算法在这些复杂交通场景中准确率平均达90.98%。
In view of the fact that the existing traffic state detection algorithm fails to cope with the complex traffic on urban roads,a new regional traffic state extraction algorithm based on Haar⁃like features and spatial⁃temporal information is proposed.Firstly,the vehicle detection algorithm based on Haar⁃like features,the edge detection method and the frame difference method are used to extract three kinds of signals of vehicles on the pavement,spatial texture and temporal texture respectively.Secondly,the three extracted signals are statistically analyzed to obtain the accurate traffic status area.The algorithm is compared with the traffic state detection algorithm based on vehicle detection and the traffic state detection algorithm on frame difference method in the complex traffic scenes with remote small targets,invisible vehicles and mixed traffic.The experimental results show that the algorithm has an accuracy rate of 90.98%in these complex traffic scenes.
作者
薛飞杨
巨永锋
宋永超
杜凯
刘维宇
XUE Feiyang;JU Yongfeng;SONG Yongchao;DU Kai;LIU Weiyu(School of Electronics and Control Engineering,Chang’an University,Xi’an 710064,China)
出处
《现代电子技术》
北大核心
2020年第1期80-85,共6页
Modern Electronics Technique
基金
国家自然科学基金资助项目（11702035）
陕西省自然科学基金资助项目（2019JQ-073）
关键词
交通状态检测
拥堵检测
车辆检测算法
空域纹理
时域纹理
信号分析
traffic state detection
congestion detection
vehicle detection algorithm
spatial texture
temporal texture
signal analysis
