基于多维度多目标的车辆计数方法  

Vehicle Counting Method based on Multi-dimensional and Multi-objective

作　　者：任元 刘战东[1] REN Yuan;LIU Zhan-dong(College of Computer Science and Technology,Xinjiang Normal University,Urumqi,Xinjiang,830054,China)

机构地区：[1]新疆师范大学计算机科学技术学院,新疆乌鲁木齐830054

出　　处：《新疆师范大学学报（自然科学版）》2023年第1期6-13,共8页Journal of Xinjiang Normal University(Natural Sciences Edition)

摘　　要：随着经济的高速发展,城市交通拥堵问题逐渐凸显,严重影响人民生活质量。为了更准确地获取视频中的车流量,文章提出了一种改进的YOLOv5+DeepSort的多维度多目标车辆计数方法。该方法可以检测、跟踪、分类车辆并对复杂路口拍摄到的视频进行计数,多维度多目标车辆计数的准确率、召回率和计数效果明显提高。对构建智能化的交通监测系统,提高交通运输效率具有重要意义。With the rapid development of economy,urban traffic congestion has become increasingly prominent,which seriously affects the improvement of people's living standards.In order to obtain the traffic flow in the video more accurately,an improved multidimensional multi-objective vehicle detection and counting method based on YOLOv5+DeepSort is proposed.This method can detect,track,classify vehicles and count videos taken at complex intersections.The accuracy,recall and counting effect of multi-dimensional multi-objective vehicle counting is significantly improved.It is of great significance to build an intelligent traffic monitoring system and improve the efficiency of transportation.

关 键 词：目标检测 车辆计数 DeepSort YOLOv5 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]