实时多车道车辆计数方法
|
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.
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.
引文
[1]TAN Ting,WANG Weifeng,WAN Jian,et al.Research on traffic flow detection system based on multi-surf placement coils[J].Journal of East China Jiaotong University,2017,34(2):60-65(in Chinese).[谭挺,王维锋,万剑,等.基于多路地感线圈的交通流量检测系统研究[J].华东交通大学学报,2017,34(2):60-65.]
[2]DENG Yi,RUAN Jingwen.The application of geomagnetic coil in the automatic detection and control system of vehicle traffic[J].Chinese Journal of Value Engineering,2014,33(5):42-43(in Chinese).[邓奕,鞠静文.地感线圈在车辆通行自动检测控制系统中的应用[J].价值工程,2014,33(5):42-43.]
[3]Nickolls J,Buck I,Garland M,et al.Scalable parallel programming with CUDA[C]//Hot Chips 20 Symposium.IEEE,2016:1-9.
[4]Jünger D,Hundt C,Domínguez JG,et al.Speed and accuracy improvement of higher-order epistasis detection on CUDA-enabled GPUs[J].Cluster Computing,2017,20(3):1899-1908.
[5]XUAN Hanyu,LIU Hongzhe,YUAN Jiazheng,et al.A robust multi-lane line detection algorithm[J].Computer Science,2017,44(11):305-313(in Chinese).[宣寒宇,刘宏哲,袁家政,等.一种鲁棒性的多车道线检测算法[J].计算机科学,2017,44(11):305-313.]
[6]CHEN Yongquan,CHEN Ying,CHEN Xuesan.Vehicle detection and tracking algorithm based on Adaboost classifier[J].Computer Technology and Development,2017,27(9):165-168(in Chinese).[陈拥权,陈影,陈学三.基于Adaboost分类器的车辆检测与跟踪算法[J].计算机技术与发展,2017,27(9):165-168.]
[7]Choudhury S,Chattopadhyay SP,Hazra TK.Vehicle detection and counting using Haar feature-based classifier[C]//Industrial Automation and Electromechanical Engineering Conference.IEEE,2017:106-109.
[8]GAO Fei,GAO Yan,XU Yunjing,et al.Multi-vehicle detection method based on ViBe and frame difference method[J].Computer Measurement&Control,2017,25(10):35-38(in Chinese).[高飞,高炎,徐云静,等.融合ViBe与帧差法的交叉路口多车辆检测方法[J].计算机测量与控制,2017,25(10):35-38.]
[9]Xia Y,Shi X,Song G,et al.Towards improving quality of video-based vehicle counting method for traffic flow estimation[J].Signal Processing,2016,120(C):672-681.
[10]Shu X,Liu L,Long X,et al.Vehicle monitoring based on taillight detection[C]//Proceedings of the International Conference on Intelligent Systems Research and Mechatronics Engineering,2015:11-13.
[11]Xu H,Zhou W,Zhu J,et al.Vehicle counting based on double virtual lines[J].Signal Image&Video Processing,2017,11(5):905-912.
[12]Wang H,Gao J,Yu L,et al.Combined improved frequencytuned with GMM algorithm for moving target detection[C]//IEEE International Conference on Mechatronics and Automation.IEEE,2017:1848-1852.
[13]Han X,Gao Y,Lu Z,et al.Research on moving object detection algorithm based on improved three frame difference method and optical flow[C]//5th International Conference on Instrumentation and Measurement,Computer,Communication and Control.IEEE,2016:580-584.
[14]ZHANG Xuan,YANG Yan.Straight line extraction algorithm based on improved random Hough transform[J].Information Technology,2017(12):6-9(in Chinese).[张炫,杨艳.基于改进随机霍夫变换的直线提取算法[J].信息技术,2017(12):6-9.]
[15]CAI Linqin,ZHENG Xuesong,WANG Niu.Target detection based on HSV color space orthogonal stereo vision[J].New Industrialization,2013,3(1):79-88(in Chinese).[蔡林沁,郑雪松,王牛.基于HSV颜色空间正交立体视觉目标检测[J].新型工业化,2013,3(1):79-88.]
[16]He L,Ren X,Gao Q,et al.The connected-component labeling problem:A review of state-of-the-art algorithms[J].Pattern Recognition,2017,70:25-43.
[2]DENG Yi,RUAN Jingwen.The application of geomagnetic coil in the automatic detection and control system of vehicle traffic[J].Chinese Journal of Value Engineering,2014,33(5):42-43(in Chinese).[邓奕,鞠静文.地感线圈在车辆通行自动检测控制系统中的应用[J].价值工程,2014,33(5):42-43.]
[3]Nickolls J,Buck I,Garland M,et al.Scalable parallel programming with CUDA[C]//Hot Chips 20 Symposium.IEEE,2016:1-9.
[4]Jünger D,Hundt C,Domínguez JG,et al.Speed and accuracy improvement of higher-order epistasis detection on CUDA-enabled GPUs[J].Cluster Computing,2017,20(3):1899-1908.
[5]XUAN Hanyu,LIU Hongzhe,YUAN Jiazheng,et al.A robust multi-lane line detection algorithm[J].Computer Science,2017,44(11):305-313(in Chinese).[宣寒宇,刘宏哲,袁家政,等.一种鲁棒性的多车道线检测算法[J].计算机科学,2017,44(11):305-313.]
[6]CHEN Yongquan,CHEN Ying,CHEN Xuesan.Vehicle detection and tracking algorithm based on Adaboost classifier[J].Computer Technology and Development,2017,27(9):165-168(in Chinese).[陈拥权,陈影,陈学三.基于Adaboost分类器的车辆检测与跟踪算法[J].计算机技术与发展,2017,27(9):165-168.]
[7]Choudhury S,Chattopadhyay SP,Hazra TK.Vehicle detection and counting using Haar feature-based classifier[C]//Industrial Automation and Electromechanical Engineering Conference.IEEE,2017:106-109.
[8]GAO Fei,GAO Yan,XU Yunjing,et al.Multi-vehicle detection method based on ViBe and frame difference method[J].Computer Measurement&Control,2017,25(10):35-38(in Chinese).[高飞,高炎,徐云静,等.融合ViBe与帧差法的交叉路口多车辆检测方法[J].计算机测量与控制,2017,25(10):35-38.]
[9]Xia Y,Shi X,Song G,et al.Towards improving quality of video-based vehicle counting method for traffic flow estimation[J].Signal Processing,2016,120(C):672-681.
[10]Shu X,Liu L,Long X,et al.Vehicle monitoring based on taillight detection[C]//Proceedings of the International Conference on Intelligent Systems Research and Mechatronics Engineering,2015:11-13.
[11]Xu H,Zhou W,Zhu J,et al.Vehicle counting based on double virtual lines[J].Signal Image&Video Processing,2017,11(5):905-912.
[12]Wang H,Gao J,Yu L,et al.Combined improved frequencytuned with GMM algorithm for moving target detection[C]//IEEE International Conference on Mechatronics and Automation.IEEE,2017:1848-1852.
[13]Han X,Gao Y,Lu Z,et al.Research on moving object detection algorithm based on improved three frame difference method and optical flow[C]//5th International Conference on Instrumentation and Measurement,Computer,Communication and Control.IEEE,2016:580-584.
[14]ZHANG Xuan,YANG Yan.Straight line extraction algorithm based on improved random Hough transform[J].Information Technology,2017(12):6-9(in Chinese).[张炫,杨艳.基于改进随机霍夫变换的直线提取算法[J].信息技术,2017(12):6-9.]
[15]CAI Linqin,ZHENG Xuesong,WANG Niu.Target detection based on HSV color space orthogonal stereo vision[J].New Industrialization,2013,3(1):79-88(in Chinese).[蔡林沁,郑雪松,王牛.基于HSV颜色空间正交立体视觉目标检测[J].新型工业化,2013,3(1):79-88.]
[16]He L,Ren X,Gao Q,et al.The connected-component labeling problem:A review of state-of-the-art algorithms[J].Pattern Recognition,2017,70:25-43.