摘要
为了对道路上行驶的车辆速度进行有效估计,提出了一种基于Harris Stephen角点检测算法和归一化互相关(Normalized Cross Correlation method,NCC)匹配算法的视频车辆检测系统。该系统对车辆进行跟踪,并测算车辆在干道和高速公路上的速度。通过使用Harris Stephen角点检测兴趣点,利用归一化互相关算法匹配对应角点,利用点对应关系确定车辆行驶的像素位移。根据车辆在连续视频帧中所有角点的总位移计算车辆的平均位移,结合帧率估计车速。实验结果表明,系统测速精度高,能实时估计出车辆行进的速度。
In order to estimate the vehicle speed on the road effectively,this paper proposes an efficient video vehicle speed measurement system based on Harris-Stephen corner detector algorithm and Normalized Cross Correlation method.The proposed system was used to track vehicle and determine vehicle speed at arterial road-ways and freeways.Harris-Stephens corner detection algorithm was used to determine interest points in the image.The normalized cross-correlation method was used to match the corresponding corner points.The displacement shift was determined in pixels corresponding to the vehicle travel.The average displacement of the vehicle was calculated according to the total displacement of the vehicle in all the corners of the continuous video frame,and the vehicle speed in the road is estimated by using the frame rate.The experimental result showed that the vehicle speed accuracy of the system is high,and the vehicle speed can be estimated in real time the system had high vehicle speed accuracy to estimate vehicle speed in real time.
作者
张俊峰
尚振宏
刘辉
ZHANG Jun-feng;SHANG Zhen-hong;LIU Hui(Faculty of Information Engineering and Automation,Kunming University of Science and Technology,Kunming 650504,China)
出处
《软件导刊》
2018年第4期127-130,134,共5页
Software Guide
关键词
车辆检测
HARRIS算法
归一化互相关
角点匹配
像素位移
车速估计
vehicle detection
Harris algorithm
normalized cross correlation method
corner matching
pixel shift
vehicle speed estimation
