智能机器人系统分析与设计
|
摘要
近年来机器人大赛在国内受到广泛的关注和欢迎,各高校和单位都以此为契机,为本校的机器人爱好者提供一个良好的动手操作平台,展现自己学校科研能力的舞台。但从整体上看,我国的机器人技术与国外先进国家还存在一定的差距,尤其是用于足球比赛的机器人,由于目前我国的视觉追踪用机器人主要还是以单片机为主,所以在运算速度和精确度上还存在一定的问题。
本文主要是通过FPGA(现场可编程门阵列)及外围配套电路来对摄像头采集的图像进行处理,根据自己的图像处理算法对每帧图像进行颜色判断,进而实现侦测目标的目的。由于FPGA在运算速度和准确度上与单片机相比具有一定的优势,其对视觉图像的处理取得了良好的效果,因此更容易被大家所认可。
在FPGA对所采集的图像进行处理分析后,再与单片机进行通讯,将处理结果传递给单片机,之后单片机再根据处理结果对目标物体的大小和距离进行判断,如果在满足形状上的要求外,被认定为是目标,则根据图像在屏幕上的位置进行距离和方位进行判断,再由单片机产生的PWM波形驱动舵机旋转以实现对目标的追踪。
由于FPGA在图像处理上具有速度上的优势,因此可以实现对每一帧画面的连续判断,这样可以通过连续的采集和连续的判断,再通过信号连续传递后实现动态性追踪。这就能够避免在踢球的过程中,由于足球被踢走后还去足球原来所在的位置去寻找的缺点。使整个寻找和追踪过程更加准确和流畅,为快速接近目标完成动作提供了可靠的保障。
本文机器人的自由度选择为18自由度,这样可以基本满足仿人的效果。可以实现人类通过各个关节的组合运动完成的行走,转向,射门等动作。使整个运动过程更能满足在比赛过程中的各种动作需要。
Robot Contest has got a lot of attention and welcome in recent years. All colleges take it as an opportunity to provide a good hands-on scientific research platform for the school's robotics enthusiasts and demonstrate the ability of scientific research of school. But as a whole, compare with advanced country our robot technology still has some gaps. Especially for soccer matches robot, as ours visual tracking robot is primarily based on singlechip, so in computing speed and precision still have problems.
In this paper, through the FPGA (Field-Programmable Gate Array) and external matching circuit to the camera collected image, and according to their own image processing algorithm to determine the color of image of each frame, so as to realize the purpose of detecting targets. As the FPGA in the computation speed and accuracy has some advantages compared to singlechip, and the good results of the visual processing, so easily approved by everyone.
Collected in the FPGA on the image processing and analysis, after that will communicate with the singlechip and handle the results passed to the singlechip, then the singlechip will deal with the results based on the target object's size and distance judgments, if the shape to meet the requirements, was identified as a goal, then based on the image on the screen of the position to judge the distance and direction, the singlechip generate PWM wave to drive steering engine to achieve the target tracking.
As the FPGA has the advantage of speed in image processing, it can achieve a continuously judge to each frame, so it can be a continuous sampling and continuous judgments, and then pass through the signal to realize dynamic nature of continuous tracking. This can avoid going to the original location to find the football after the ball has been kicked out during the game. This make sure that the entire process of finding and tracking become more accurately and smoothly, that provides reliable protection for fast close to the target.
This paper adopts 18 degrees of freedom, so that can basically meet the imitation. It can achieve walking, turning, shooting etc by a combination of various joints. To meet the various actions required in the contest.
本文主要是通过FPGA(现场可编程门阵列)及外围配套电路来对摄像头采集的图像进行处理,根据自己的图像处理算法对每帧图像进行颜色判断,进而实现侦测目标的目的。由于FPGA在运算速度和准确度上与单片机相比具有一定的优势,其对视觉图像的处理取得了良好的效果,因此更容易被大家所认可。
在FPGA对所采集的图像进行处理分析后,再与单片机进行通讯,将处理结果传递给单片机,之后单片机再根据处理结果对目标物体的大小和距离进行判断,如果在满足形状上的要求外,被认定为是目标,则根据图像在屏幕上的位置进行距离和方位进行判断,再由单片机产生的PWM波形驱动舵机旋转以实现对目标的追踪。
由于FPGA在图像处理上具有速度上的优势,因此可以实现对每一帧画面的连续判断,这样可以通过连续的采集和连续的判断,再通过信号连续传递后实现动态性追踪。这就能够避免在踢球的过程中,由于足球被踢走后还去足球原来所在的位置去寻找的缺点。使整个寻找和追踪过程更加准确和流畅,为快速接近目标完成动作提供了可靠的保障。
本文机器人的自由度选择为18自由度,这样可以基本满足仿人的效果。可以实现人类通过各个关节的组合运动完成的行走,转向,射门等动作。使整个运动过程更能满足在比赛过程中的各种动作需要。
Robot Contest has got a lot of attention and welcome in recent years. All colleges take it as an opportunity to provide a good hands-on scientific research platform for the school's robotics enthusiasts and demonstrate the ability of scientific research of school. But as a whole, compare with advanced country our robot technology still has some gaps. Especially for soccer matches robot, as ours visual tracking robot is primarily based on singlechip, so in computing speed and precision still have problems.
In this paper, through the FPGA (Field-Programmable Gate Array) and external matching circuit to the camera collected image, and according to their own image processing algorithm to determine the color of image of each frame, so as to realize the purpose of detecting targets. As the FPGA in the computation speed and accuracy has some advantages compared to singlechip, and the good results of the visual processing, so easily approved by everyone.
Collected in the FPGA on the image processing and analysis, after that will communicate with the singlechip and handle the results passed to the singlechip, then the singlechip will deal with the results based on the target object's size and distance judgments, if the shape to meet the requirements, was identified as a goal, then based on the image on the screen of the position to judge the distance and direction, the singlechip generate PWM wave to drive steering engine to achieve the target tracking.
As the FPGA has the advantage of speed in image processing, it can achieve a continuously judge to each frame, so it can be a continuous sampling and continuous judgments, and then pass through the signal to realize dynamic nature of continuous tracking. This can avoid going to the original location to find the football after the ball has been kicked out during the game. This make sure that the entire process of finding and tracking become more accurately and smoothly, that provides reliable protection for fast close to the target.
This paper adopts 18 degrees of freedom, so that can basically meet the imitation. It can achieve walking, turning, shooting etc by a combination of various joints. To meet the various actions required in the contest.
引文
[1]谢涛,徐建峰,张永学等.仿人机器人的研究历史现状及展望.机器人,2002,Vol.24:5～20.
[2] Massato H. Humanoid Robot. Journal of the Robotics Society of Japan.1997,15(7):983～985.
[3] Kazuo H.Current and Future Perspective of Honda Humanoid Robot.IEEE/RSJInternational Conference on Intelligent Robots and System.1997:500～508.
[4] Kazuo H,Massato H,Yuji H.The Development of Honda Humanoid Robot. Procedings of the 1998 IEEE International Conference on Robotics&Automation. Leuven, Belgium, May,1998:1322～1326.
[5]李实,徐旭明,叶棒.国际机器人足球比赛及其相关技术.机器人,2000,22(5) :420～426.
[6]杨林权,罗忠文.机器人足球竞赛与设计技术.湖北:华中科技大学出版社,2009.
[7]张庆玲,杨勇. FPGA原理与实践.北京:北京航空航天大学出版社,2006.
[8]叶淦华. FPGA嵌入式应用系统开发典型实例.北京:中国电力出版社,2005.
[9]景晓军.图像处理技术及其应用.北京:国防工业出版社,2005.
[10]张洪刚,陈光,郭军编.图像处理与识别.北京:北京邮电大学出版社,2006.
[11]魏博宇,李科杰,高峻峣等.移动机器人视觉系统与摄像机标定.微计算机信息, 2009,25(8):171～172.
[12]汪倩倩.基于双目视觉的跟踪机器人系统的设计.微计算机信息, 2005,25(8):74~76.
[13]李金良,孙友霞,包继华等.救援机器人目标跟踪控制的研究工矿自动化2009,(12):22～24.
[14]杜鑫峰,熊蓉,褚健.仿人足球机器人视觉系统快速识别与精确定位.浙江大学学报, 2009,43(11): 1975～1981.
[15] DU Xin ZHAO. Robot Tracking by Color Image. Geo-spatial Information Science,2007, 10(1),33~36.
[16]龚剑,方康玲.机器人足球视觉目标的快速识别.微计算机信息,2007,23(3):248~250.
[17]刘欢,魏立峰,王健.机器人视觉伺服系统的标定.微计算机信息,2007,23(4):278~279.
[18]刘小力,田梦倩,罗翔等机器人视觉伺服系统的图像处理和标定技术.中国制造业信息化, 2005,34(10):119~121.
[19]杨莉,杜艳红,隋金雪等.一种基于颜色特征的目标识别算法.微计算机信息, 2007,23(5):195~197.
[20]费峥峰,赵雁南.一种快速彩色图像颜色分割算法微计算机信息,2007,23(10):185~187.
[21]关军,杨明.基于视觉的高速寻线机器人设计与实现.微计算机信息,2007,23(11):240~242.
[22]王建平,张涛.移动机器人检测与跟踪系统的研究与设计浙江大学学报, 2007,41(10):1710~1714.
[23]胡伟华,李文锋,张帆.基于ARM9的智能教育机器人设计.微计算机信息, 2009,25(12):189～190.
[24]史庆军,李敏,姜重然,陈文平,徐斌山仿人机器人关节驱动系统设计微计算机信息, 2009,25(12):196～198.
[25]晁阳.单片机MCS-51原理及应用开发教程.北京:清华大学出版社,2007.
[26]马宏绪,张彭,张良起.两足步行机器人准动态步行姿态稳定性及姿态控制.机器人, 19(3):180～186.
[27]赵景山.机器人机构自有度分析理论.北京:科学出版社,2009.
[28]颜嘉男.伺服电机应用技术.北京:科学出版社,2008.
[29]谢维成杨加国.单片机原理与应用及C51程序设计.北京:清华大学出版社,2009.
[30]蒋海林,张国良,宋海涛.基于ARM9的嵌入式足球机器人运动控制系统微计算机信息, 2009,25(8):97～99.
[31]张诌,黄强,李光日等.小型仿人机器人控制系统设计.微计算机信息,2007,23(8):1~3.
[32]陈宇.基于单片机智能机器人控制系统研究设计.微计算机信息,2007,23(12):243~244.
[33]曾国强,葛良全.机器人比赛中直流电机驱动电路的设计微计算机信息,2008,24:236~238.
[34]钟秋波,潘启树,洪炳镕等.一种仿人机器人复杂运动控制方法.华中科技大学学报, 2008,36:199~202.
[35]朴松昊,孙立宁,钟秋波等.仿人机器人视觉颜色空间的图像分割算法.华中科技大学学报, 2008,36:98~101.
[36]李瑞峰,赵立军,范旭丰.室内环境下视觉目标跟踪.华中科技大学学报,2008,36:149~151.
[37]王晓荣,张印强,陆艳.全自主移动足球机器人目标识别.微计算机信息, 2008,24(10):226~227.
[38]白素华,魏立峰,吕薇.基于颜色相似度量的足球机器人目标识别.机器人技术, 2008,24(10):247~248.
[39]张一鸣,秦世引.基于单目视觉的移动机器人测距方法.机器人技术,2008,24(10):224~225.
[2] Massato H. Humanoid Robot. Journal of the Robotics Society of Japan.1997,15(7):983～985.
[3] Kazuo H.Current and Future Perspective of Honda Humanoid Robot.IEEE/RSJInternational Conference on Intelligent Robots and System.1997:500～508.
[4] Kazuo H,Massato H,Yuji H.The Development of Honda Humanoid Robot. Procedings of the 1998 IEEE International Conference on Robotics&Automation. Leuven, Belgium, May,1998:1322～1326.
[5]李实,徐旭明,叶棒.国际机器人足球比赛及其相关技术.机器人,2000,22(5) :420～426.
[6]杨林权,罗忠文.机器人足球竞赛与设计技术.湖北:华中科技大学出版社,2009.
[7]张庆玲,杨勇. FPGA原理与实践.北京:北京航空航天大学出版社,2006.
[8]叶淦华. FPGA嵌入式应用系统开发典型实例.北京:中国电力出版社,2005.
[9]景晓军.图像处理技术及其应用.北京:国防工业出版社,2005.
[10]张洪刚,陈光,郭军编.图像处理与识别.北京:北京邮电大学出版社,2006.
[11]魏博宇,李科杰,高峻峣等.移动机器人视觉系统与摄像机标定.微计算机信息, 2009,25(8):171～172.
[12]汪倩倩.基于双目视觉的跟踪机器人系统的设计.微计算机信息, 2005,25(8):74~76.
[13]李金良,孙友霞,包继华等.救援机器人目标跟踪控制的研究工矿自动化2009,(12):22～24.
[14]杜鑫峰,熊蓉,褚健.仿人足球机器人视觉系统快速识别与精确定位.浙江大学学报, 2009,43(11): 1975～1981.
[15] DU Xin ZHAO. Robot Tracking by Color Image. Geo-spatial Information Science,2007, 10(1),33~36.
[16]龚剑,方康玲.机器人足球视觉目标的快速识别.微计算机信息,2007,23(3):248~250.
[17]刘欢,魏立峰,王健.机器人视觉伺服系统的标定.微计算机信息,2007,23(4):278~279.
[18]刘小力,田梦倩,罗翔等机器人视觉伺服系统的图像处理和标定技术.中国制造业信息化, 2005,34(10):119~121.
[19]杨莉,杜艳红,隋金雪等.一种基于颜色特征的目标识别算法.微计算机信息, 2007,23(5):195~197.
[20]费峥峰,赵雁南.一种快速彩色图像颜色分割算法微计算机信息,2007,23(10):185~187.
[21]关军,杨明.基于视觉的高速寻线机器人设计与实现.微计算机信息,2007,23(11):240~242.
[22]王建平,张涛.移动机器人检测与跟踪系统的研究与设计浙江大学学报, 2007,41(10):1710~1714.
[23]胡伟华,李文锋,张帆.基于ARM9的智能教育机器人设计.微计算机信息, 2009,25(12):189～190.
[24]史庆军,李敏,姜重然,陈文平,徐斌山仿人机器人关节驱动系统设计微计算机信息, 2009,25(12):196～198.
[25]晁阳.单片机MCS-51原理及应用开发教程.北京:清华大学出版社,2007.
[26]马宏绪,张彭,张良起.两足步行机器人准动态步行姿态稳定性及姿态控制.机器人, 19(3):180～186.
[27]赵景山.机器人机构自有度分析理论.北京:科学出版社,2009.
[28]颜嘉男.伺服电机应用技术.北京:科学出版社,2008.
[29]谢维成杨加国.单片机原理与应用及C51程序设计.北京:清华大学出版社,2009.
[30]蒋海林,张国良,宋海涛.基于ARM9的嵌入式足球机器人运动控制系统微计算机信息, 2009,25(8):97～99.
[31]张诌,黄强,李光日等.小型仿人机器人控制系统设计.微计算机信息,2007,23(8):1~3.
[32]陈宇.基于单片机智能机器人控制系统研究设计.微计算机信息,2007,23(12):243~244.
[33]曾国强,葛良全.机器人比赛中直流电机驱动电路的设计微计算机信息,2008,24:236~238.
[34]钟秋波,潘启树,洪炳镕等.一种仿人机器人复杂运动控制方法.华中科技大学学报, 2008,36:199~202.
[35]朴松昊,孙立宁,钟秋波等.仿人机器人视觉颜色空间的图像分割算法.华中科技大学学报, 2008,36:98~101.
[36]李瑞峰,赵立军,范旭丰.室内环境下视觉目标跟踪.华中科技大学学报,2008,36:149~151.
[37]王晓荣,张印强,陆艳.全自主移动足球机器人目标识别.微计算机信息, 2008,24(10):226~227.
[38]白素华,魏立峰,吕薇.基于颜色相似度量的足球机器人目标识别.机器人技术, 2008,24(10):247~248.
[39]张一鸣,秦世引.基于单目视觉的移动机器人测距方法.机器人技术,2008,24(10):224~225.