实时多车道车辆计数方法  

Real-time multi-lane vehicle counting method

作　　者：王超[1] 陈庆奎[1] WANG Chao;CHEN Qing-kui(College of Optical-Electrical and Computer Engineering,University of Shanghai for Science and Technology,Shanghai 200093,China)

机构地区：[1]上海理工大学光电信息与计算机工程学院,上海200093

出　　处：《计算机工程与设计》2019年第5期1383-1389,共7页Computer Engineering and Design

基　　金：国家自然科学基金项目(61572325;60970012);高等学校博士学科点专项科研博导基金项目(20113120110008);上海重点科技攻关基金项目(14511107902;16DZ1203603);上海市工程中心建设基金项目(GCZX14014);上海智能家居大规模物联共性技术工程中心基金项目(GCZX14014);上海市一流学科建设基金项目(XTKX2012);沪江基金研究基地专项基金项目(C14001)

摘　　要：为智能检测多车道车辆数目,提出一种实时多车道车辆计数方法。利用运动车辆完成整个车道区域提取排除非车道区域的干扰,在道路背景重建得到无车辆遮挡的道路背景后完成车道线检测和拟合,得到多车道区域;在此基础上,通过提取车尾灯红色区域并建立相应的匹配规则匹配成对车尾灯,解决车辆并排同速问题,完成车辆计数的任务,实现多车道车辆计数。在CUDA平台下使用图像处理器(GPU)NVIDIA GTX680显卡对算法进行加速,可以达到28ms/帧的处理速率,验证了算法的实时性。To intelligently detect the number of multi-lane vehicles,a real-time multi-lane vehicle counting method was proposed.The entire lane area was extracted using motion vehicles to eliminate the interference of the non-lane area,and after reconstructing the road background without vehicle obstruction,the lane line was detected and fitted to obtain a multi-lane area.Based on lane division,the red taillights were extracted and the corresponding matching rules were established to match the pair taillights,and the problem of side-by-side vehicles with same speed was solved,the vehicle counting task was completed,and the multi-lane vehicle counting was achieved.NVIDIA GTX680 GPU (graphic processing unit) was used to accelerate the algorithm on CUDA platform and the processing speed of 28 ms per frame was achieved,which verified the real-time performance of the algorithm.

关 键 词：智能检测 多车道区域 背景重建 匹配规则 统一计算架构 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]