摘要
提出了一种统计监控视频中行人流量的方法。文中方法利用背景差法提取运动物体,并对运动物体进行跟踪。提取物体的HOG特征,并利用线性SVM判决提取的HOG特征是否符合行人要求,以达到统计视频中行人流量的目的。实验表明在人流量不是特别密集的场景,可以准确地统计视频中的行人数量。
A new method for counting pedestrian flow based on video monitoring system was proposed in this paper. This method can adaptively update the background, extracting moving objects using background difference method,and tracking moving objects. Finally extracting HOG features of the object,using Linear SVM judges the HOG feature if it meets the requirements of pedestrians or not,in order to achieve the purpose of counting pedestrian flow. The experiments show that this method can accurately statistics the number of pedestrian if the traffic was not particularly congestion.
出处
《信息技术》
2015年第8期22-25,共4页
Information Technology
基金
江苏省自然科学基金(BK20130473)
镇江市社会发展科技支撑项目(SH2013015)
