一种新型多车道车流量检测算法  

New Algorithm of Multilane Traffic Flow Detection

作　　者：何清[1] 陈海进[1] 施慧华[1] 

机构地区：[1]南通大学专用集成电路设计重点实验室,江苏南通226019

出　　处：《电视技术》2014年第19期192-195,210,共5页Video Engineering

基　　金：江苏省高校自然科学研究重大项目(12KJA51002);南通大学自然科学研究项目(11ZJ001;12B031)

摘　　要：为了实时有效地检测道路路口车流量信息,并为交通控制和管理提供准确的交通流数据,提出了一种新型的车流量检测算法。通过利用现有的路面标记进行图像对称分割并计算图像灰度值方差,来判断有无车辆通过,进而实现车流量计算。仿真结果表明,该算法不仅简单,易于实现,而且检测准确率高,实时性好,能够有效地为智能交通灯控制提供信息数据。In order to detect traffic flow effectively in real-time and provide accurate traffic flow data for traffic control and management, a new de- tection algorithm is presented. The algorithm uses existing road markings to determine whether vehicle has passed by calculating the gray value vari- ance of symmetrical image, then realizes vehicle detection. Simulation results showed that the algorithm is not only simple and easy to implement, but also is of high detection accuracy. Furthermore, it has good real-time performance, which can provide information for intelligent traffic light control ef- fectively.

关 键 词：车流量检测 图像灰度值方差 智能交通 视频检测 

分 类 号：TN911.73[电子电信—通信与信息系统] TP391.41[电子电信—信息与通信工程]