驾驶员疲劳检测方法研究及嵌入式实现
|
摘要
由于疲劳驾驶引发的交通事故日渐增多,疲劳驾驶的检测已成为交通管理部门亟待解决的问题之一。如何有效的检测驾驶员的疲劳驾驶状态,已经成为汽车安全技术领域的一个研究热点。本文研究了驾驶员疲劳检测相关算法,并以美国德州仪器(TI)公司生产的TMS320DM6437高性能数字视频开发板为开发平台,设计并实现了驾驶员疲劳检测系统。
目前许多研究者集中于通过跟踪驾驶员的人脸,眼睛,瞳孔等,得到头部转动和方向,眼睑运动,眨眼频率,注意力方向等监测驾驶员疲劳驾驶或精神分散状态。本文的驾驶员疲劳检测系统,采用一个车载CCD摄像头采集图像,在高性能数字视频开发板TMS320DM6437 EVM上对图像进行实时处理,运用图像处理的相关算法对每一帧图像进行分析,实现人脸定位,继而实现人眼定位,最后在借鉴目前广泛采用的PERCLOS算法基础上,设计出一种驾驶员疲劳检测算法。实验表明,本系统定位眼睛准确,能准确的判定疲劳状态。
本文使用MathWorks公司提供的Matlab和Simulink进行算法仿真。利用DSP模块,定点模块集以及图像和视频处理模块集等模块构建算法模型,直接生成硬件平台可执行的代码,将代码下载到开发平台上进行测试。由于这种仿真可直接在硬件环境中进行,避免了通用计算机到硬件平台移植过程中出现的问题,加快了开发过程。
本文研究了TMS320DM6437数字视频开发评估板,对其硬件各个模块的性能和系统视频设备驱动进行了分析,并对其数据传输通道进行配置；对驾驶员疲劳检测算法进行研究,完成人脸定位,人眼定位以及驾驶员疲劳状态的判定,算法完成后进行仿真验证；完成视频设备的软件驱动设计,将驾驶员疲劳检测算法在TMS320DM6437数字视频开发评估板上实现,从而最终完成驾驶员疲劳检测系统。
As the increasing number of traffic accidents due to drivers'fatigue, the detection of drivers'fatigue has become one of the most important subjects in the vehicle management department. How to detect the drivers'fatigue conditions effectively has become a hot topic in the field of vehicle safety. In this paper, the algorithms related with the detection of drivers'fatigue are studied. A detection system on drivers'fatigue is designed and implemented by using the platform of TMS320DM6437 which is produced by TI Co.U.S.
Now, many researchers have focused on monitoring the driver's face, eye, pupil and so on to obtain his/her face rotation and orientation, eye activities, eye blinking rate, gaze direction, finally to determine his/her fatigue or distraction state. This drivers' fatigue system acquires the images through a CCD camera of vehicles and do the real-time processing on TMS320DM6437 EVM which is a high-efficiency digital video developing board. Each frame of images is analyzed using the image processing algorithms to achieve location of faces and eyes. Finally, based on PERCLOS algorithm a method is designed to detect driver's fatigue. Experimental results show that this system can locate the eyes and determine fatigue condition accurately.
This paper uses the tools of MATLAB and Simulink provided by Math Works Company to simulate algorithms. Algorithm model is designed by DSP modules, the sets of the fixed-point modules and image and video processing modules. The codes are generated directly which execute on hardware platform and downloaded to the developing platform for testing. As this simulation can be carried out directly on the hardware platform, it can avoid the problems which turn up in the transplant process between common computers and hardware platform, and speed up the developing process.
This paper has researched on TMS320DM6437 digital video developing evaluation board, analyzed the performances of each hardware modules and systemic video devices'drivers and configured their data transmission channels; this paper has also studied on the algorithm of drivers'fatigue detection, which incompletes the functions of the location of faces and eyes and determining the state of drivers'fatigue, the algorithms are simulated and turned out after completion; this paper has made general designs of video devices'software drivers, the detecting algorithm on drivers'fatigue is achieved on the TMS320DM6437 digital video developing evaluation board, and ultimately complete the drivers'fatigue detecting system.
目前许多研究者集中于通过跟踪驾驶员的人脸,眼睛,瞳孔等,得到头部转动和方向,眼睑运动,眨眼频率,注意力方向等监测驾驶员疲劳驾驶或精神分散状态。本文的驾驶员疲劳检测系统,采用一个车载CCD摄像头采集图像,在高性能数字视频开发板TMS320DM6437 EVM上对图像进行实时处理,运用图像处理的相关算法对每一帧图像进行分析,实现人脸定位,继而实现人眼定位,最后在借鉴目前广泛采用的PERCLOS算法基础上,设计出一种驾驶员疲劳检测算法。实验表明,本系统定位眼睛准确,能准确的判定疲劳状态。
本文使用MathWorks公司提供的Matlab和Simulink进行算法仿真。利用DSP模块,定点模块集以及图像和视频处理模块集等模块构建算法模型,直接生成硬件平台可执行的代码,将代码下载到开发平台上进行测试。由于这种仿真可直接在硬件环境中进行,避免了通用计算机到硬件平台移植过程中出现的问题,加快了开发过程。
本文研究了TMS320DM6437数字视频开发评估板,对其硬件各个模块的性能和系统视频设备驱动进行了分析,并对其数据传输通道进行配置；对驾驶员疲劳检测算法进行研究,完成人脸定位,人眼定位以及驾驶员疲劳状态的判定,算法完成后进行仿真验证；完成视频设备的软件驱动设计,将驾驶员疲劳检测算法在TMS320DM6437数字视频开发评估板上实现,从而最终完成驾驶员疲劳检测系统。
As the increasing number of traffic accidents due to drivers'fatigue, the detection of drivers'fatigue has become one of the most important subjects in the vehicle management department. How to detect the drivers'fatigue conditions effectively has become a hot topic in the field of vehicle safety. In this paper, the algorithms related with the detection of drivers'fatigue are studied. A detection system on drivers'fatigue is designed and implemented by using the platform of TMS320DM6437 which is produced by TI Co.U.S.
Now, many researchers have focused on monitoring the driver's face, eye, pupil and so on to obtain his/her face rotation and orientation, eye activities, eye blinking rate, gaze direction, finally to determine his/her fatigue or distraction state. This drivers' fatigue system acquires the images through a CCD camera of vehicles and do the real-time processing on TMS320DM6437 EVM which is a high-efficiency digital video developing board. Each frame of images is analyzed using the image processing algorithms to achieve location of faces and eyes. Finally, based on PERCLOS algorithm a method is designed to detect driver's fatigue. Experimental results show that this system can locate the eyes and determine fatigue condition accurately.
This paper uses the tools of MATLAB and Simulink provided by Math Works Company to simulate algorithms. Algorithm model is designed by DSP modules, the sets of the fixed-point modules and image and video processing modules. The codes are generated directly which execute on hardware platform and downloaded to the developing platform for testing. As this simulation can be carried out directly on the hardware platform, it can avoid the problems which turn up in the transplant process between common computers and hardware platform, and speed up the developing process.
This paper has researched on TMS320DM6437 digital video developing evaluation board, analyzed the performances of each hardware modules and systemic video devices'drivers and configured their data transmission channels; this paper has also studied on the algorithm of drivers'fatigue detection, which incompletes the functions of the location of faces and eyes and determining the state of drivers'fatigue, the algorithms are simulated and turned out after completion; this paper has made general designs of video devices'software drivers, the detecting algorithm on drivers'fatigue is achieved on the TMS320DM6437 digital video developing evaluation board, and ultimately complete the drivers'fatigue detecting system.
引文
[1]张伦维.疲劳驾驶的克星-防疲劳驾驶系统让驾驶更安全.当代汽车,2008.
[2]2009年全国道路交通事故概况,道路交通管理,2009(1)：6-8.
[3]张灵聪,王正国,朱佩芳,等.汽车驾驶疲劳研究综述.人类工效应,2003,9(1)：39-42.
[4]Lal S. K. L., Craig A Driver fatigue:Electroencephalography and psychological assessment. Psychophysiology,2006(39):313-321.
[5]A kerstedt T.,Torsvall L, Gillberg M. Sleepiness and shiftwork. filed studies,1998:95-106.
[6]Santamaria J. Chiappa K.H.The EEG of drowsiness in normal adults. Clin Neurophysiol,1997(4):327-382.
[7]Yabuta K., Iizuka H., Yanagishima T., Kataoka Y. The development of drowsiness warning devices. US Department of Transportation,1985, Section 4:282.
[8]Grace R., Byrne V. E. D. M. Bierman. A drowsy driver detection system for heavy vehicles, Proceedings of 17th Digital Avionics Systems Conference, Bellevue, WA, USA,1998, (136)1-(136)8.
[9]Dinges D. F., Grace R. PERCLOS:a valid psychophysiological measure of alertness as assessed by psychomotor vigilance. Federal Highway Administration, Office of Motor Carriers,1998.
[10]李治春,何仁,林谋有.驾驶员疲劳检测预警技术研究现状及发展趋势.第四届亚太可持续发展交通与环境技术大会,中国西安,2005.
[11]Qiang Ji, Zhiwei Zhu, Peilin Lan. Real-Time Nonintrusive Monitoring and Prediction of Driver Fatigue[J]. IEEE Transactions on Vehicular Technology,2004,53(4).
[12]Tim H, Laurence H, Gerald P K et al. Fatigue Detection Technologies for Drivers:A Review of Existing Operator-Centred Systems. IEE 2001 Conference Publication, 2001,481.
[13]王雪立.基于机器视觉的嵌入式驾驶疲劳检测系统研究：(硕十学位论文).北京：首都师范大学,2006.
[14]王光娟。基于DSP的驾驶疲劳检测系统的研究与实现：(硕十学位论文).镇江：江苏大学,2007.
[15]黎亚平,周杰,黄磊,吴仲兵.国内外驾驶疲劳状态检测技术的现状与发展.南京：中国单片机公共实验室南京研发中心,2008.
[16]Texas Instruments Incoroprated. TMS320DM6437 Evaluation Module Technical Reference. Dallas:Texas Instruments Incoroprated,2006,12-13.
[17]Texas Instruments Incoroprated. TMS320DM64x DSP Two level memory Reference Guide. Dallas:Texas Instruments Incoroprated,2006,11-18.
[18]Texas Instruments Incoroprated. MS320C64x/C64x+ DSP CPU and Instruction Set Reference Guide (Rev. H). Dallas:Texas Instruments Incoroprated,2008,31-38.
[19]Texas Instruments Incorporated. TVP5146 NTSC/PAL/SECAM 4x10-Bit Digital Video Decoder With MacrovisionTM Detection, YCbCr/RGB Inputs,5-Lines Comb Filter and SCART Support. Dallas:Texas Instruments Incorporated,2007, SLES084C:1-53.
[20]Texas Instruments Incoroprated. MS320DM643x DMP Peripherals Overview Reference Guide (Rev. A). Dallas:Texas Instruments Incoroprated,2008,31-38.
[21]Texas Instruments Incorporated. TMS320DM6643x DMP Video Processing Front End(VPFE) User's Guide.Dallas:Texas Intruments Incoporated,2008, SPRU977A:14-18.
[22]彭启宗.达芬奇技术—数字图像／视频信号处理新平台.北京：电子工业出版,2008：134-160.
[23]Texas Instruments Incorporated著.王军宁,何迪,马娟译.TI DSP/BIOS用户手册与驱动开发。北京：清华大学出版社,2007.
[24]李方慧,王飞,何佩琨等.TMS320C6000系列DSP原理与应用.北京：电子工业出版社,2003.
[25]臧博,吴裕斌,曹丹华.基于GIO/FVID 的 DSP视频处理驱动程序.武汉：华中科技大学,2006.
[26]Texas Instruments Incorporated. The TMS320DM642 Video Port Mini-Driver. Dallas:Texas Instruments Incorporated,2003, SPRA918A:3-23.
[27]刘向宇.DSP嵌入式常用模块与综合系统设计实例精讲.北京：电子工业出版社,2009：135-159.
[28]Texas Instruments Incorporated. How to Use the EDMA3 Driver on a TMS320DM643x Device. Dallas:Texas Instruments Incorporated,2008, SPRAAN4A3:1-20.
[29]王良民,张建明,詹永照等.人脸检测研究现状和发展.江苏大学学报,2003,24(3).
[30]李刚,高政.人脸检测技术研究与发展。计算机与现代文化,2003,4：7-8.
[31]张蓉蓉.人脸检测技术的研究(硕十学位论文).北京：北京邮电大学,2006.
[32]陆宗骐,金登男.Visual C++. net图像处理编程。北京：清华大学出版社,2006：120-190.
[33]张宏林.精通Visual C++数字图像处理典型算法及实现.北京：人民邮电出版社,2006：162-198.
[34]何建东,耿楠,张义宽.数字图像处理.西安：西安电子科技大学出版社,2003：164-234.
[35]贾云得.机器视觉.北京：科学出版社,2000：15-20.
[36]http://baike. baidu. com/view/841383. htm?fr=ala0_1.
[37]马野伦.基于数字图像处理的人眼定位算法研究(硕士论文).吉林：吉林大学,2006.
[38]王荣本,郭克友,储江伟.适用驾驶员疲劳状态监测的人眼定位方法研究.公路交通科技,2003：111-114.
[39]冈萨雷斯R C著.阮秋琦译.数字图像处理学(第二版).北京：电子工业出版社,2001：375-407.
[40]夏芹,宋义伟,朱学峰.基于P E R C L O S的驾驶疲劳监控方法进展.自动化技术与应用,2008(11)：81-83.
[41]Dinges D F, Grace R. PERCLOS:A Valid Psycho physiological Measure of Alertness as Assessed by Psychomotor vigilance[R]. Washington:Federal Highway Administration, Office of Motor Carriers,1998.
[42]陶芬.全天侯疲劳驾驶监测系统的研究及实现(硕士论文).南京：南京理工大学,2009.
[43]郑培.机动车驾驶员疲劳测评方法的研究(博士论文).北京：中国农业大学,2002.
[44]黄永安,马路,刘慧敏.MATALAB7.0/Simulink6.0建模仿真开发与高级工程应用.北京：清华大学出版社,2005,210-256.
[45]薛定宇,陈阳泉.基于MATLAB/Simulink的系统仿真技术与应用.北京：清华大学出版社,2002,34-60.
[46]李真芳,苏涛,黄小宇.DSP程序开发-MATLAB调试及直接目标代码生成.西安：西安电子科技大学出版社,2003：294-325.
[47]赵家祥.DSP系统设计和BIOS编程及应用实例-基于TMS320C67x系列DSP芯片.北京：机械工业出版社,2007.
[48]郭森楙,颜允圣著,贾洪峰译.数字信号处理器-体系结构、实现与应用.北京：清华大学出版社,2005.
[49]Texas Instruments Incoroprated. TMS320 DSP/BIOS User's Guide. Dallas:Texas Instruments,2004, SPRU423F:22-23.
[50]Andrew Bateman lain Paterson-Stephens著,陈健,陈伟,汪书宁等译.DSP算法、应用与设计.北京：机械工业出版社,2003：15-35.
[51]徐勇.基于DSP的驾驶员疲劳监测方法研究：(硕士学位论文).南京：南京理工大学,2006.
[52]Texas Instruments Incorporated. Code Engine Application Developer user's Guide. Dallas:Texas Instruments Incorporated,2007, SPRUE67D:3-23.
[53]赵勇.DAVINCI技术原理应用指南.南京：东南大学出版社,2008.
[54]Texas Instruments Incorporated. Programming Details of Codec engine for DaVincTM technology. Dallas:Texas Instruments Incorporated,2006, SPRY091:19-47.
[55]徐盛,胡建凌.TM320DSP算法标准。北京：清华大学出版社,2007：7-10.
[2]2009年全国道路交通事故概况,道路交通管理,2009(1)：6-8.
[3]张灵聪,王正国,朱佩芳,等.汽车驾驶疲劳研究综述.人类工效应,2003,9(1)：39-42.
[4]Lal S. K. L., Craig A Driver fatigue:Electroencephalography and psychological assessment. Psychophysiology,2006(39):313-321.
[5]A kerstedt T.,Torsvall L, Gillberg M. Sleepiness and shiftwork. filed studies,1998:95-106.
[6]Santamaria J. Chiappa K.H.The EEG of drowsiness in normal adults. Clin Neurophysiol,1997(4):327-382.
[7]Yabuta K., Iizuka H., Yanagishima T., Kataoka Y. The development of drowsiness warning devices. US Department of Transportation,1985, Section 4:282.
[8]Grace R., Byrne V. E. D. M. Bierman. A drowsy driver detection system for heavy vehicles, Proceedings of 17th Digital Avionics Systems Conference, Bellevue, WA, USA,1998, (136)1-(136)8.
[9]Dinges D. F., Grace R. PERCLOS:a valid psychophysiological measure of alertness as assessed by psychomotor vigilance. Federal Highway Administration, Office of Motor Carriers,1998.
[10]李治春,何仁,林谋有.驾驶员疲劳检测预警技术研究现状及发展趋势.第四届亚太可持续发展交通与环境技术大会,中国西安,2005.
[11]Qiang Ji, Zhiwei Zhu, Peilin Lan. Real-Time Nonintrusive Monitoring and Prediction of Driver Fatigue[J]. IEEE Transactions on Vehicular Technology,2004,53(4).
[12]Tim H, Laurence H, Gerald P K et al. Fatigue Detection Technologies for Drivers:A Review of Existing Operator-Centred Systems. IEE 2001 Conference Publication, 2001,481.
[13]王雪立.基于机器视觉的嵌入式驾驶疲劳检测系统研究：(硕十学位论文).北京：首都师范大学,2006.
[14]王光娟。基于DSP的驾驶疲劳检测系统的研究与实现：(硕十学位论文).镇江：江苏大学,2007.
[15]黎亚平,周杰,黄磊,吴仲兵.国内外驾驶疲劳状态检测技术的现状与发展.南京：中国单片机公共实验室南京研发中心,2008.
[16]Texas Instruments Incoroprated. TMS320DM6437 Evaluation Module Technical Reference. Dallas:Texas Instruments Incoroprated,2006,12-13.
[17]Texas Instruments Incoroprated. TMS320DM64x DSP Two level memory Reference Guide. Dallas:Texas Instruments Incoroprated,2006,11-18.
[18]Texas Instruments Incoroprated. MS320C64x/C64x+ DSP CPU and Instruction Set Reference Guide (Rev. H). Dallas:Texas Instruments Incoroprated,2008,31-38.
[19]Texas Instruments Incorporated. TVP5146 NTSC/PAL/SECAM 4x10-Bit Digital Video Decoder With MacrovisionTM Detection, YCbCr/RGB Inputs,5-Lines Comb Filter and SCART Support. Dallas:Texas Instruments Incorporated,2007, SLES084C:1-53.
[20]Texas Instruments Incoroprated. MS320DM643x DMP Peripherals Overview Reference Guide (Rev. A). Dallas:Texas Instruments Incoroprated,2008,31-38.
[21]Texas Instruments Incorporated. TMS320DM6643x DMP Video Processing Front End(VPFE) User's Guide.Dallas:Texas Intruments Incoporated,2008, SPRU977A:14-18.
[22]彭启宗.达芬奇技术—数字图像／视频信号处理新平台.北京：电子工业出版,2008：134-160.
[23]Texas Instruments Incorporated著.王军宁,何迪,马娟译.TI DSP/BIOS用户手册与驱动开发。北京：清华大学出版社,2007.
[24]李方慧,王飞,何佩琨等.TMS320C6000系列DSP原理与应用.北京：电子工业出版社,2003.
[25]臧博,吴裕斌,曹丹华.基于GIO/FVID 的 DSP视频处理驱动程序.武汉：华中科技大学,2006.
[26]Texas Instruments Incorporated. The TMS320DM642 Video Port Mini-Driver. Dallas:Texas Instruments Incorporated,2003, SPRA918A:3-23.
[27]刘向宇.DSP嵌入式常用模块与综合系统设计实例精讲.北京：电子工业出版社,2009：135-159.
[28]Texas Instruments Incorporated. How to Use the EDMA3 Driver on a TMS320DM643x Device. Dallas:Texas Instruments Incorporated,2008, SPRAAN4A3:1-20.
[29]王良民,张建明,詹永照等.人脸检测研究现状和发展.江苏大学学报,2003,24(3).
[30]李刚,高政.人脸检测技术研究与发展。计算机与现代文化,2003,4：7-8.
[31]张蓉蓉.人脸检测技术的研究(硕十学位论文).北京：北京邮电大学,2006.
[32]陆宗骐,金登男.Visual C++. net图像处理编程。北京：清华大学出版社,2006：120-190.
[33]张宏林.精通Visual C++数字图像处理典型算法及实现.北京：人民邮电出版社,2006：162-198.
[34]何建东,耿楠,张义宽.数字图像处理.西安：西安电子科技大学出版社,2003：164-234.
[35]贾云得.机器视觉.北京：科学出版社,2000：15-20.
[36]http://baike. baidu. com/view/841383. htm?fr=ala0_1.
[37]马野伦.基于数字图像处理的人眼定位算法研究(硕士论文).吉林：吉林大学,2006.
[38]王荣本,郭克友,储江伟.适用驾驶员疲劳状态监测的人眼定位方法研究.公路交通科技,2003：111-114.
[39]冈萨雷斯R C著.阮秋琦译.数字图像处理学(第二版).北京：电子工业出版社,2001：375-407.
[40]夏芹,宋义伟,朱学峰.基于P E R C L O S的驾驶疲劳监控方法进展.自动化技术与应用,2008(11)：81-83.
[41]Dinges D F, Grace R. PERCLOS:A Valid Psycho physiological Measure of Alertness as Assessed by Psychomotor vigilance[R]. Washington:Federal Highway Administration, Office of Motor Carriers,1998.
[42]陶芬.全天侯疲劳驾驶监测系统的研究及实现(硕士论文).南京：南京理工大学,2009.
[43]郑培.机动车驾驶员疲劳测评方法的研究(博士论文).北京：中国农业大学,2002.
[44]黄永安,马路,刘慧敏.MATALAB7.0/Simulink6.0建模仿真开发与高级工程应用.北京：清华大学出版社,2005,210-256.
[45]薛定宇,陈阳泉.基于MATLAB/Simulink的系统仿真技术与应用.北京：清华大学出版社,2002,34-60.
[46]李真芳,苏涛,黄小宇.DSP程序开发-MATLAB调试及直接目标代码生成.西安：西安电子科技大学出版社,2003：294-325.
[47]赵家祥.DSP系统设计和BIOS编程及应用实例-基于TMS320C67x系列DSP芯片.北京：机械工业出版社,2007.
[48]郭森楙,颜允圣著,贾洪峰译.数字信号处理器-体系结构、实现与应用.北京：清华大学出版社,2005.
[49]Texas Instruments Incoroprated. TMS320 DSP/BIOS User's Guide. Dallas:Texas Instruments,2004, SPRU423F:22-23.
[50]Andrew Bateman lain Paterson-Stephens著,陈健,陈伟,汪书宁等译.DSP算法、应用与设计.北京：机械工业出版社,2003：15-35.
[51]徐勇.基于DSP的驾驶员疲劳监测方法研究：(硕士学位论文).南京：南京理工大学,2006.
[52]Texas Instruments Incorporated. Code Engine Application Developer user's Guide. Dallas:Texas Instruments Incorporated,2007, SPRUE67D:3-23.
[53]赵勇.DAVINCI技术原理应用指南.南京：东南大学出版社,2008.
[54]Texas Instruments Incorporated. Programming Details of Codec engine for DaVincTM technology. Dallas:Texas Instruments Incorporated,2006, SPRY091:19-47.
[55]徐盛,胡建凌.TM320DSP算法标准。北京：清华大学出版社,2007：7-10.