基于DSP机器视觉道路划线车定位系统的研究与设计
|
摘要
基于我国目前道路划线车的现状，研究和开发智能化、高精度、稳定性好的划线车势在必行。而划线系统的快速精度定位，是开发智能化道路画线系统的关键。因此本文主要研究基于DSP机器视觉道路划线定位系统，主要的研究工作包括：
1．研究图像预处理算法以及定位特征参数获取算法。在对图像进行前期处理过程中，采用图像的灰度化、平滑去噪、二值化处理、多次闭运算、边缘提取算法等组合，然后根据这些算法组合，对图像进行预处理，以利于特征参数的精确提取。另外，通过摄像机的标定、小孔成像原理、直线投影规律、Hough变换方法研究，获得道路边缘直线坐标参数。
2．设计系统的软硬件平台。在硬件设计中，给出了主要的硬件电路图：视频采集显示模块、EMIF内存扩展模块、电源时钟复位模块、DSP通信模块等。软件设计包括应用软件和系统软件，应用软件负责实现系统的所用应用功能，而系统软件负责初始化硬件平台，为应用软件准备好系统环境。在系统软件设计中，定制了适合DSP硬件环境的操作系统DSP／BIOS，编写了视频采集显示驱动以及PCI驱动程序。在PC端，开发了基于WDM模型的PCI驱动程序。在应用软件设计中，DSP端，采用了TI公司的RF5参考框架构建应用程序；PC端，构建基于MFC的控制界面应用程序。
最后的实验结果显示，本文提出的系统设计方案，无论是前期的算法组合方案还是软硬件平台方案，均达到了定位系统研究的要求，为进一步研究应用打下了坚实的基础。
It is necessary to study and develop the lineation vehicle with intellectualization, high accuracy and good stability in the light of the present situation of the drawing vehicle of our country. The key technology is how the vehicle lineation system is located accurately and quickly. Therefore this study focuses on the positioning-system of the road-lineation based on machine vision. The main researching works is as following:
1. The algorithms of image pre-processing were discussed and the characteristic parameters of positioning were determined. Image-gradation, smoothing, binarization "close" and withdrawing the edge algorithms were discussed and their combination was employed in the image pre-processing for the accurately extraction of characteristic parameters. In addition, the line coordinate parameters of the road-edge were obtained based on the calibration of camera, principle of pinhole model, the rule of linear projection and Hough transformation.
2. The platform consisting of software and hardware was designed. In the design of hardware, the main circuits of hardware: the module of capturing and displaying, the module of EMIF memory expansion, the module of power, clock and reset, the module of DSP correspondence and so on were developed. The design of software includes the software of application and system. The application software is to realize the application function of the system, whereas the system software to initialize the system platform and prepare the environment of the system for the application software. DSP/BIOS of the operating system, which is adapt to the dsp-hardware environment, was customized in the process of designing system software. The drive programs about capturing and displaying device and PCI device were developed. The drive program of PCI based on the module of WDM was also written. In the design of the application program, RF5 of TI Corporation was used to develop the application program of the DSP and construct the control interface application program based on MFC.
Experimental result showed that both algorithm combination and platform of software and hardware could meet the need of the preliminary research of this positioning system. These acquired achievements have built the solid foundation for the future application research.
1．研究图像预处理算法以及定位特征参数获取算法。在对图像进行前期处理过程中，采用图像的灰度化、平滑去噪、二值化处理、多次闭运算、边缘提取算法等组合，然后根据这些算法组合，对图像进行预处理，以利于特征参数的精确提取。另外，通过摄像机的标定、小孔成像原理、直线投影规律、Hough变换方法研究，获得道路边缘直线坐标参数。
2．设计系统的软硬件平台。在硬件设计中，给出了主要的硬件电路图：视频采集显示模块、EMIF内存扩展模块、电源时钟复位模块、DSP通信模块等。软件设计包括应用软件和系统软件，应用软件负责实现系统的所用应用功能，而系统软件负责初始化硬件平台，为应用软件准备好系统环境。在系统软件设计中，定制了适合DSP硬件环境的操作系统DSP／BIOS，编写了视频采集显示驱动以及PCI驱动程序。在PC端，开发了基于WDM模型的PCI驱动程序。在应用软件设计中，DSP端，采用了TI公司的RF5参考框架构建应用程序；PC端，构建基于MFC的控制界面应用程序。
最后的实验结果显示，本文提出的系统设计方案，无论是前期的算法组合方案还是软硬件平台方案，均达到了定位系统研究的要求，为进一步研究应用打下了坚实的基础。
It is necessary to study and develop the lineation vehicle with intellectualization, high accuracy and good stability in the light of the present situation of the drawing vehicle of our country. The key technology is how the vehicle lineation system is located accurately and quickly. Therefore this study focuses on the positioning-system of the road-lineation based on machine vision. The main researching works is as following:
1. The algorithms of image pre-processing were discussed and the characteristic parameters of positioning were determined. Image-gradation, smoothing, binarization "close" and withdrawing the edge algorithms were discussed and their combination was employed in the image pre-processing for the accurately extraction of characteristic parameters. In addition, the line coordinate parameters of the road-edge were obtained based on the calibration of camera, principle of pinhole model, the rule of linear projection and Hough transformation.
2. The platform consisting of software and hardware was designed. In the design of hardware, the main circuits of hardware: the module of capturing and displaying, the module of EMIF memory expansion, the module of power, clock and reset, the module of DSP correspondence and so on were developed. The design of software includes the software of application and system. The application software is to realize the application function of the system, whereas the system software to initialize the system platform and prepare the environment of the system for the application software. DSP/BIOS of the operating system, which is adapt to the dsp-hardware environment, was customized in the process of designing system software. The drive programs about capturing and displaying device and PCI device were developed. The drive program of PCI based on the module of WDM was also written. In the design of the application program, RF5 of TI Corporation was used to develop the application program of the DSP and construct the control interface application program based on MFC.
Experimental result showed that both algorithm combination and platform of software and hardware could meet the need of the preliminary research of this positioning system. These acquired achievements have built the solid foundation for the future application research.
引文
[1] 刘湘晨，薛龙，戴静君．公路划线机的应用现状和发展方向[J]．北京石油化工学院学报，2004(9)，12(3)：62～64
[2] 沈松云．国外划线机[J]．建筑机械，1998(10)：38～39
[3] 屈耀庭．道路交通标线质量和划线机的选择[J]．筑路机械与施工机械化，1997(1)，14(66)：23～24
[4] 筑养路机械的发展趋势及市场前景[J]，中国土建工程信息网(http://www.tujian.com)
[5] 何超．机器人视觉系统中的图像分割与目标检测[J]．浙江大学学报，2003．2～33
[6] 段峰．基于机器视觉的智能空瓶检测机器人研究[J]．长春理工大学学报，2002．7～42
[7] Haralick R M. Digital step edges from zero crossing of second directional derivatives .IEEE Trans. on PAMI, 1984, 6(1): 58～68
[8] Marr D, Hildreth E. Theory of Edge Detection. London: Proc Roy Soc, 1980
[9] Canny J. A Computational approach to edge detection. IEEE Trans on PAMI, 1986, 8(6): 679～698
[10] 刘教民，赵小英，魏世泽，孙壮志．TMS320C40实现图像高速采集与处理系统[J]．河北科技大学学报，2001(3)，22(58)：1～5
[11] 陈益新，陈思培，李雯，张桂林．基于多个DSP的图像处理SIMD并行结构[J]．华中理工大学学报，1997(3)：30～32
[12] 任俊，涂晓昱，傅一平，袁丁，李志能．利用DSP的实时图像识别系统的设计与应用[J]．光电工程，2004(2)：66～69
[13] 徐婉莹，刘建军，黄新生．基于CPLD和DSP的高速图像采集技术研究[J]．电子工程师，2004(6)，30(6)：48～50
[14] Lu S, et al. Improving edge measurement on noisy image by hierachical NN. Pattern Recognition Letters, 1991: 12～155
[15] 杨淑莹．VC++图像处理程序设计[M]．北京：清华大学出版社．北京交通大学出版社．2005．1
[16] 陆宗骐．C／C++图像处理编程[M]．北京：清华大学出版社．2005．1
[17] 夏良正．数字图像处理[M]．江苏：东南大学出版社．2003．8
[18] 郎锐．数字图像处理学Visual C++实现[M]．北京：北京希望电子出版社．2003．1
[19] 马颂德，张正友．计算机视觉一计算理论与实践[M]．北京：科学出版社．1998
[20] Yongtae Do. Applications of neural networks for stereo-Camera Calibration. Proc. of International Joint Conference on Neural Networks. Washington: IEEE, 1999. 2179～2722
[21] Ganapathy S. Decomposition of transformation matrices for robot vision. Proc. Int. Conference on Robotics and Automation. 1984. 130～139
[22] Abdel-Aziz Y. I., Karara H. M.. Direct linear transformation from comparator coordinates into object space coordinates in close-range photogrammetry. In U. of Illinois Symposium on Close-Range Photogrammetry. Urbana: Univ. of Illinois at Urbana-Champaign, 1971. 1～18
[23] 吴文琪，孙增圻．机器视觉中的摄像机标定方法综述[J]．计算机应用研究，2004(2)：4～6
[24] 吴福朝，李华，胡占义．基于主动视觉系统的摄像机自标定方法研究[J]．自动化学报，2001(11)，27(6)：752～762
[25] 周俊．农用轮式移动机器人视觉导航系统的研究．南京农业大学博士学位论文，2003
[26] Yi Ma, Jana Kosecka, Shankar S Sastry. Vision guided navigation for a nonholonomic mobile robot. IEEE Transaction on Robotics and Automation, 1999,15(3):521～536
[27] 张雄伟，陈亮，徐光辉．DSP芯片的原理与开发应用[M]．北京：电子工业出版社．2005．10
[28] 戴逸民，梁晓雯，裴小平．基于DSP的现代电子系统设计[M]．北京：电子工业出版社．2002．12
[29] 江思敏，刘畅．TMS320C6000DSP应用开发教程[M]．北京：机械工业出版社．2005．2
[30] 2006年SEED DSP+FPGA培训讲义
[31] 张勇．C／C++语言硬件程序设计[M]．陕西：西安电子科技大学出版社．2003．5
[32] 李强．重庆大学．硕士学位论文．基于DSP的智能车辆机器视觉系统研究，2004．6
[33] 陈景岩．吉林大学．硕士学位论文．基于DM642的视频通信系统，2006．4
[34] 张国华，郑成芳，张鹏．基于PCI总线的实时DSP图像处理平台设计[J]．电光与控制，2005(6)，12(3)：66～69
[35] PCI2040 PCI-DSP Bridge Controller Data Book, SCPS048, Texas Instruments Inc, 1999
[36] 靖文广 吕汉兴．基于PCI总线的嵌入式实时DSP图像采集系统[J]．电子技术应用，2003(12)：16～18
[37] 熊智华，李强，张颖．基于TMS320VC5416的图像采集处理系统[J]．计算机工程与应用，2005(4)：109～111
[38] SEED-VPM642用户指南：TMS320DM642视频DSP处理模板．2005．4
[39] TMS320C6000 Peripheral Component Interconnect (PCI) Reference Guide. SPRU581, Texas Instruments Inc, 2004.6
[40] TMS320C6000 Host Port Interface (HPI) Reference Guide. SPRU578, Texas Instruments Inc, 2004.6
[41] TMS320C6000 Ethernet Media Access Controller(EMAC)/Management Data Input/Output(MDIO) Module Reference Guide. SPRU628, Texas Instruments Inc, 2004.6
[42] Code Composer Studio User' s Guide. SPRU328B, Texas Instruments Inc, 2000.2
[43] TMS320 DSP/BIOS User' s Guide. SPRU423B, Texas Instruments Inc, 2002.11
[44] DSP/BIOS Driver Developer' s Guide. SPRU616, Texas Instruments Inc, 2002.11
[45] Reference Frameworks For eXpressDSP Software:RF5, An Extensive High-Density System. SPRA795A, Texas Instruments Inc, 2003.4
[46] The TMS320DM642 Video Port Mini-Driver. SPRA918A, Texas Instruments Inc, 2003.4
[47] A DSP/BIOS PCI Device Driver for TMS320C64xx DSPs. SPRA845A, Texas Instruments Inc, 2003.6
[48] Tools Tech Note2 Getting Started with DM642 PCI. Spectrum Digital. 2003.6
[49] SEED PCI(WDM)驱动程序接口函数及使用说明．VERSION 1．1
[50] 武安河．Windows2000／XP WDM设备驱动程序开发[M]．北京：电子工业出版社．2005．5
[51] 唐克．MFC程序设计(Using Visual C++ 6.0)[M]．北京：希望电子出版社．2002．10
[52] 陈建春．Visual C++高级编程技术(开发实例剖析)[M]．北京：电子工业出版社．2000．4
[2] 沈松云．国外划线机[J]．建筑机械，1998(10)：38～39
[3] 屈耀庭．道路交通标线质量和划线机的选择[J]．筑路机械与施工机械化，1997(1)，14(66)：23～24
[4] 筑养路机械的发展趋势及市场前景[J]，中国土建工程信息网(http://www.tujian.com)
[5] 何超．机器人视觉系统中的图像分割与目标检测[J]．浙江大学学报，2003．2～33
[6] 段峰．基于机器视觉的智能空瓶检测机器人研究[J]．长春理工大学学报，2002．7～42
[7] Haralick R M. Digital step edges from zero crossing of second directional derivatives .IEEE Trans. on PAMI, 1984, 6(1): 58～68
[8] Marr D, Hildreth E. Theory of Edge Detection. London: Proc Roy Soc, 1980
[9] Canny J. A Computational approach to edge detection. IEEE Trans on PAMI, 1986, 8(6): 679～698
[10] 刘教民，赵小英，魏世泽，孙壮志．TMS320C40实现图像高速采集与处理系统[J]．河北科技大学学报，2001(3)，22(58)：1～5
[11] 陈益新，陈思培，李雯，张桂林．基于多个DSP的图像处理SIMD并行结构[J]．华中理工大学学报，1997(3)：30～32
[12] 任俊，涂晓昱，傅一平，袁丁，李志能．利用DSP的实时图像识别系统的设计与应用[J]．光电工程，2004(2)：66～69
[13] 徐婉莹，刘建军，黄新生．基于CPLD和DSP的高速图像采集技术研究[J]．电子工程师，2004(6)，30(6)：48～50
[14] Lu S, et al. Improving edge measurement on noisy image by hierachical NN. Pattern Recognition Letters, 1991: 12～155
[15] 杨淑莹．VC++图像处理程序设计[M]．北京：清华大学出版社．北京交通大学出版社．2005．1
[16] 陆宗骐．C／C++图像处理编程[M]．北京：清华大学出版社．2005．1
[17] 夏良正．数字图像处理[M]．江苏：东南大学出版社．2003．8
[18] 郎锐．数字图像处理学Visual C++实现[M]．北京：北京希望电子出版社．2003．1
[19] 马颂德，张正友．计算机视觉一计算理论与实践[M]．北京：科学出版社．1998
[20] Yongtae Do. Applications of neural networks for stereo-Camera Calibration. Proc. of International Joint Conference on Neural Networks. Washington: IEEE, 1999. 2179～2722
[21] Ganapathy S. Decomposition of transformation matrices for robot vision. Proc. Int. Conference on Robotics and Automation. 1984. 130～139
[22] Abdel-Aziz Y. I., Karara H. M.. Direct linear transformation from comparator coordinates into object space coordinates in close-range photogrammetry. In U. of Illinois Symposium on Close-Range Photogrammetry. Urbana: Univ. of Illinois at Urbana-Champaign, 1971. 1～18
[23] 吴文琪，孙增圻．机器视觉中的摄像机标定方法综述[J]．计算机应用研究，2004(2)：4～6
[24] 吴福朝，李华，胡占义．基于主动视觉系统的摄像机自标定方法研究[J]．自动化学报，2001(11)，27(6)：752～762
[25] 周俊．农用轮式移动机器人视觉导航系统的研究．南京农业大学博士学位论文，2003
[26] Yi Ma, Jana Kosecka, Shankar S Sastry. Vision guided navigation for a nonholonomic mobile robot. IEEE Transaction on Robotics and Automation, 1999,15(3):521～536
[27] 张雄伟，陈亮，徐光辉．DSP芯片的原理与开发应用[M]．北京：电子工业出版社．2005．10
[28] 戴逸民，梁晓雯，裴小平．基于DSP的现代电子系统设计[M]．北京：电子工业出版社．2002．12
[29] 江思敏，刘畅．TMS320C6000DSP应用开发教程[M]．北京：机械工业出版社．2005．2
[30] 2006年SEED DSP+FPGA培训讲义
[31] 张勇．C／C++语言硬件程序设计[M]．陕西：西安电子科技大学出版社．2003．5
[32] 李强．重庆大学．硕士学位论文．基于DSP的智能车辆机器视觉系统研究，2004．6
[33] 陈景岩．吉林大学．硕士学位论文．基于DM642的视频通信系统，2006．4
[34] 张国华，郑成芳，张鹏．基于PCI总线的实时DSP图像处理平台设计[J]．电光与控制，2005(6)，12(3)：66～69
[35] PCI2040 PCI-DSP Bridge Controller Data Book, SCPS048, Texas Instruments Inc, 1999
[36] 靖文广 吕汉兴．基于PCI总线的嵌入式实时DSP图像采集系统[J]．电子技术应用，2003(12)：16～18
[37] 熊智华，李强，张颖．基于TMS320VC5416的图像采集处理系统[J]．计算机工程与应用，2005(4)：109～111
[38] SEED-VPM642用户指南：TMS320DM642视频DSP处理模板．2005．4
[39] TMS320C6000 Peripheral Component Interconnect (PCI) Reference Guide. SPRU581, Texas Instruments Inc, 2004.6
[40] TMS320C6000 Host Port Interface (HPI) Reference Guide. SPRU578, Texas Instruments Inc, 2004.6
[41] TMS320C6000 Ethernet Media Access Controller(EMAC)/Management Data Input/Output(MDIO) Module Reference Guide. SPRU628, Texas Instruments Inc, 2004.6
[42] Code Composer Studio User' s Guide. SPRU328B, Texas Instruments Inc, 2000.2
[43] TMS320 DSP/BIOS User' s Guide. SPRU423B, Texas Instruments Inc, 2002.11
[44] DSP/BIOS Driver Developer' s Guide. SPRU616, Texas Instruments Inc, 2002.11
[45] Reference Frameworks For eXpressDSP Software:RF5, An Extensive High-Density System. SPRA795A, Texas Instruments Inc, 2003.4
[46] The TMS320DM642 Video Port Mini-Driver. SPRA918A, Texas Instruments Inc, 2003.4
[47] A DSP/BIOS PCI Device Driver for TMS320C64xx DSPs. SPRA845A, Texas Instruments Inc, 2003.6
[48] Tools Tech Note2 Getting Started with DM642 PCI. Spectrum Digital. 2003.6
[49] SEED PCI(WDM)驱动程序接口函数及使用说明．VERSION 1．1
[50] 武安河．Windows2000／XP WDM设备驱动程序开发[M]．北京：电子工业出版社．2005．5
[51] 唐克．MFC程序设计(Using Visual C++ 6.0)[M]．北京：希望电子出版社．2002．10
[52] 陈建春．Visual C++高级编程技术(开发实例剖析)[M]．北京：电子工业出版社．2000．4