摘要
视频检测的研究是交通参数采集的主要问题,核心是对运动目标的检测。为克服传统基于灰度的背景帧差法所存在的缺陷,提出了建立在YCbCr彩色空间的背景帧差法检测运动目标,分割阈值采用基于迭代的最佳阈值的确定方法,并提出了梯度极值法得到最清晰、完整的车辆进行车流量的统计。经仿真结果表明提出的车辆检测算法不仅能够取得很高的检测精度,而且算法简单、运行速度快。
Video based Vehicle detection is an important way for collecting traffic parameters.Its core is motion object detection.In order to overcome the shortcomings of traditional background difference based on gray,a background difference based on YCbCr color space is proposed to detect motion object and iteration operation to determine the threshold of partition image.The paper proposes grad maximum to find the clearest and most integrated vehicle image,which provides the basis for vehicle counting.The simulation results demonstrate that the proposed algorithm not only can improve detection accuracy,but also is very simple and has high running speed.
出处
《计算机仿真》
CSCD
北大核心
2010年第1期285-287,308,共4页
Computer Simulation
关键词
彩色空间
梯度极值法
车辆检测
背景帧差
Color space
Grad extremum
Vehicle detection
Background difference
