基于FPGA的小型组足球机器人视觉系统研究与设计
|
摘要
机器人世界杯是一个世界性的机器人足球比赛,其鼓励人工智能及相关领域的研究。而视觉系统作为足球机器人的主要感知设备,其性能的好坏对机器人性能的发挥起着决定性作用。
随着机器视觉的作用越来越显著,人们对图像处理的研究越来越深入。由于小型组足球机器人视觉系统对识别的准确性,实时性,以及对现场环境的适应性的要求,使得它是一个很好的研究机器视觉的平台。
本项目的目的是开发一个供湖北工业大学的足球机器人队使用的视觉系统的工作原型。因为现存的小型组足球机器人的视觉系统通常是利用计算机对摄入的图像进行处理,计算机一边要进行图像的数据采集,一边要对图像进行各种算法的处理,图像的处理所需的运算,大量的占用了CPU的时间,从而影响了识别响应的实时性,并且由于足球机器人快速的运动和激烈的对抗,要求视频捕获频率为30帧/秒,即必须在16ms内完成视觉信息处理、通讯、动作规划等一系列任务,以形成下一周期的比赛策略,这对视觉系统的实时性提出了更高的要求,一些时间复杂度高的算法和大量占用CPU时间的方法很难直接使用。因此本文提出了基于FPGA的视觉系统。本文以FPGA为平台,使用VHDL硬件描述语言设计并实现了一种基于YCbCr色彩空间的算法。该系统不仅满足了视觉系统对实时性、稳定性和准确性的需求,还为系统提供了灵活性,可以根据需要对系统内部逻辑功能进行配置。不仅如此,基于FPGA的视觉系统大大缩小了视觉系统的规模,更有助于系统的稳定、携带和调试。
实验结果证明所设计的小型组足球机器人视觉系统满足了实时性、准确性、灵活性和稳定性的要求,具有较好的应用前景。
The RoboCup is an international robot soccer competition, which encourages the study of artificial intelligence and related fields. The soccer robot vision system as the main perception equipment, its performance is good or bad performance of the robot to play a decisive role to play.
With the growing role of machine vision significantly, people are more and more in-depth study of image processing. As a small team league of soccer robot vision system for recognition accuracy, real-time, as well as the adaptability of on-site environmental requirements, making it a good platform for machine vision research.
The purpose of this project is to develop one for the Hubei University of Technology team of soccer robots using a working prototype of the visual system. Because the existing small group of soccer robot vision system is usually using computer processing image.when computer to conduct collection of the image data, it processes various algorithms to image, image processing required for computing a large number of takes up CPU resources and time, thus affecting the recognition of real-time response.And because soccer robot rapid movement and intense confrontation, require video capture frequency of 30 frames/s, which must be completed within 16ms of visual information processing, communication, action planning and a series of tasks to form the next cycle race strategy, which the real-time visual system put forward higher requirements for some time a high complexity algorithm and CPU-time is very difficult to direct performce.
Therefore, this article proposed FPGA-based vision system. In this paper, as a platform FPGA, using VHDL language was designed and implemented an algorithm based on the YCbCr color space. The system not only meets the real-time vision system, stability and accuracy of demand, but also provide flexibility for the system may be based on the need configure logic functions meeting the demand within the system. Moreover, FPGA-based vision system has significantly reduced the size of the visual system, but also contribute to system stability, carry and debugging.
At last,we test the vision system by RoboCup small team league.Experiments show that the feasibility of soccer robot vision system satisfies the request of real-time, precision, flexibility and stability.It has a bright future in application prospects.
随着机器视觉的作用越来越显著,人们对图像处理的研究越来越深入。由于小型组足球机器人视觉系统对识别的准确性,实时性,以及对现场环境的适应性的要求,使得它是一个很好的研究机器视觉的平台。
本项目的目的是开发一个供湖北工业大学的足球机器人队使用的视觉系统的工作原型。因为现存的小型组足球机器人的视觉系统通常是利用计算机对摄入的图像进行处理,计算机一边要进行图像的数据采集,一边要对图像进行各种算法的处理,图像的处理所需的运算,大量的占用了CPU的时间,从而影响了识别响应的实时性,并且由于足球机器人快速的运动和激烈的对抗,要求视频捕获频率为30帧/秒,即必须在16ms内完成视觉信息处理、通讯、动作规划等一系列任务,以形成下一周期的比赛策略,这对视觉系统的实时性提出了更高的要求,一些时间复杂度高的算法和大量占用CPU时间的方法很难直接使用。因此本文提出了基于FPGA的视觉系统。本文以FPGA为平台,使用VHDL硬件描述语言设计并实现了一种基于YCbCr色彩空间的算法。该系统不仅满足了视觉系统对实时性、稳定性和准确性的需求,还为系统提供了灵活性,可以根据需要对系统内部逻辑功能进行配置。不仅如此,基于FPGA的视觉系统大大缩小了视觉系统的规模,更有助于系统的稳定、携带和调试。
实验结果证明所设计的小型组足球机器人视觉系统满足了实时性、准确性、灵活性和稳定性的要求,具有较好的应用前景。
The RoboCup is an international robot soccer competition, which encourages the study of artificial intelligence and related fields. The soccer robot vision system as the main perception equipment, its performance is good or bad performance of the robot to play a decisive role to play.
With the growing role of machine vision significantly, people are more and more in-depth study of image processing. As a small team league of soccer robot vision system for recognition accuracy, real-time, as well as the adaptability of on-site environmental requirements, making it a good platform for machine vision research.
The purpose of this project is to develop one for the Hubei University of Technology team of soccer robots using a working prototype of the visual system. Because the existing small group of soccer robot vision system is usually using computer processing image.when computer to conduct collection of the image data, it processes various algorithms to image, image processing required for computing a large number of takes up CPU resources and time, thus affecting the recognition of real-time response.And because soccer robot rapid movement and intense confrontation, require video capture frequency of 30 frames/s, which must be completed within 16ms of visual information processing, communication, action planning and a series of tasks to form the next cycle race strategy, which the real-time visual system put forward higher requirements for some time a high complexity algorithm and CPU-time is very difficult to direct performce.
Therefore, this article proposed FPGA-based vision system. In this paper, as a platform FPGA, using VHDL language was designed and implemented an algorithm based on the YCbCr color space. The system not only meets the real-time vision system, stability and accuracy of demand, but also provide flexibility for the system may be based on the need configure logic functions meeting the demand within the system. Moreover, FPGA-based vision system has significantly reduced the size of the visual system, but also contribute to system stability, carry and debugging.
At last,we test the vision system by RoboCup small team league.Experiments show that the feasibility of soccer robot vision system satisfies the request of real-time, precision, flexibility and stability.It has a bright future in application prospects.
引文
[1]Fira Organization Official Home page http://www. fira. net[z]
[2]RoboCup Organization Official Home page http://www.robocup.org[Z]
[3]The F180 Technical Committee. http://small-size.informatik. uni-bremen. de/rules:main This is an electronic document. Data retrieved:July 20,2009. Laws of the F180 League 2009, Law 2-The Ball and Law 4-The Robotic Equipment,2009, pages 5 and 7.
[4]刘作时,林桂娟,张海英.足球机器人的视觉系统的研制[J].传感器技术,2004,23(3)：38-40.
[5]郭成果.在不稳定光照条件下足球机器人视觉系统的设计[J].机器人：2006-02-12.
[6]谭然.光强自适应的足球机器人视觉系统的设计研究[D].西北工业大学,2007：13-16
[7]付昱强.基于FPGA的图像处理算法的研究与硬件设计,南昌大学硕士论文,2006： 58页
[8]董兵.FPGA图像处理板设计,哈尔滨工程大学硕士论文,2007：4页
[9]Z.-K. Huang and D.-H. Liu, Segmentation of color image using EM algorithm in HSV color space, [J] in Proceedings of IEEE International Conference on Information Acquisition, Jul.2007, pp.316-319.
[10]许志倩,张作龙,朱万胜.足球机器人视觉系统目标识别算法的改进,流体传动与控制,总第22期,第2页。
[11]陈景航.MiroSot足球机器人视觉子系统的改进算法研究.广东工业大学硕士学位论文.2006：20-24页.
[12]Studio encoding parameters of digital television for standard 4:3 and wide-screen 16:9 aspect ratios. http://inst.eecs.berkeley.edu/-cs150/Documents/ITU601. PDF This is an electronic document. Data retrieved:July 25,2009.
[13]Interface for digital component video signals in 525-line and 625-line television operating at the 4:2:2 level of recommendation ITU-R BT.601 (Part A). http://inst.eecs.berkeley.edu/-cs150/Documents/ITU656.PDF This is an electronic document. Data retrieved:July 25,2009.
[14]Yang Liu, Smail Menani, and Johan Dams. Botnia Team Description Paper for RoboCup 2007 Small Size League [R] 2007
[15]Yanowitz S D, Bruckstein A M. A new method for image segmentation. Computer Vision, Graphics, and Image Processing,1989,46:82-95
[16]周鲜成.图像分割方法及其应用研究综述[J]信息技术2007.12
[17]Bruce J, Balch T, Veloso M. Fast and Inexpensive Color Image Segmentation for Interactive Robots [A]. Proceedings of the 2000 IEEE/RSJ International Conference on Intelligent Robots and Systems[C]. USA:IEEE,2000.2061-2066
[18]James Bruce. Realtime machine vision perception and prediction. Undergraduate thesis, Camegie Mellon University May 2000
[19]Burce J , Veloso M. Fast and accurate vision-based Pattern detection and Identification[A]. Proceedings of the 2003 IEEE International Conference on Robotics and Automation[C].USA:IEEE,2003.1277-1282.
[20]William O. McCalister, Chih-Cheng Hung. Image Segmentation Using Run-Length Coding Technique[C].SCIA 2003,LNCS 2749,343-348
[21]严学强,叶秀清,基于量化图像直方图的最大熵阈值处理算法,[J]：模式识别与人工智能,1998.3
[22]周德龙,潘泉,最大熵阈值处理算法,[J]：软件学报,2001.9
[23]Dah-Jye Lee, James Archibald, Xiaoqian Xu and Pengcheng Zhan. Using distance transform to solve real-time machine vision inspection problems. SpringerLink.2006.
[24]Maria E. Angelopoulou, Christos-Savvas Bouganis, Peter Y. K. Cheung and George A. Constantinides. FPGA-Based Real-Time Super-Resolution on an Adaptive Image Sensor. SpringerLink.2006.
[25]李铁军ROBOCUP中型组机器人视觉系统的研究与开发山东大学硕士论文2006年11月
[26]Liang M, Sun Z K. A neural network model for image segmentation. ICAI&NN92, Switzerland.
[27]张建岭,刘军,基于小波变换和形态学的图像分割方法,[J]中国人民公安大学学报(自然科学版),2007.02
[28]胡慧之FPGA在图像处理器中的应用《信息技术》2006年31卷8余楠,孙芳,基于改进遗传算法的图像分割,[J]：计算机与数字工程,2007.8
[29]朱志刚,等译.数字图像处理[M].北京：电子工业出版社,1998
[30]曹爱增,石兴广,刘峰,等.机器人足球颜色分类研究[J].济南大学学报(自然科学版),2005,19(3)：239-241.
[31]Kenneth R. Castleman著,朱志刚等译.数字图象处理[M].北京：电子工业出版社.2002.
[32]Rafael C. Gonzalez, Richard E. Woods,阮秋琦阮宇智等译.数字图象处理(第二版)[M].北京：电子工业出版社.2003.
[33]Yasuda, G. Bin Ge, Color based object recognition, localization, and movement control for multiple wheeled mobile robots, [J]Industrial Electronics Society, 2004. IECON 2004.30th Annual Conference of IEEE
[34]Parag H. Batavia, Sanjiv Singh. Obstacle Detection Using Adaptive Color Segmentation and Color Stereo Homography. Robotics and Automation,2001. Proceedings 2001 ICRA. IEEE International Conference on
[35]黄晶,赵臣,周明明.基于快速彩色空间变换的足球机器人目标搜索[J].哈尔滨工业大学学报.2003.Vol 35 No.9：1036-1039.
[36]杜啸晓,杨新,施鹏飞.机器人足球视党系统的软件研究[J]数据采集与处理.2001第16卷第2期：160-164.
[37]赵小松,张宏伟,张国雄,李真.摄像机标定技术的研究[J].机械工程学报,2002,38(3)：149-151
[38]吴文琪,孙增圻.机器视觉中的摄像机定标方法综述[J]计算机应用研究2004.2
[40]韩娟娟邓文怡娄小平 基于FPGA的图像增强处理系统的设计与实现《微计算机信息》2007年26期起止页码：229-230,122
[41]Faig,W., Calibration of Close-Range Photogrammetry System:Mathematical Formulation, [J], No.12, December 1975,pp.1479-1486.
[42]Wilfried Kubinger, Franz Rinnerthaler, Christoph Sulzbachner, Josef Langer and Martin Humenberger. An Embedded Vision Sensor for Robot Soccer. SpringerLink.2007.
[43]Z. Zhang and C. Loop, Estimating the Fundamental Matrix by Transforming Image Points in Projective Space, Computer Vision and Image Understanding, Vol.82, No.2, pages 174-180,2001
[44]Zhengyou Zhang, Flexible Camera Calibration By Viewing a Plane From Unknown Orientations [J]. Microsoft Research, One Microsoft Way, Redmond, WA 98052-6399, USA. Computer Vision, IEEE International Conference on,vol.1.
[45]杨泽辉孙运强姚爱琴基于FPGA的图像增强压缩系统的设计 《黑龙江科技信息》2007年10期起止页码：85-85
[46]Frederic Devernay, Olivier Faugeras. Straight lines have to be straight Automatic calibration and removal of distortion form scenes of structured environments[J]. Machine Vision and Applications,2001,13:14-24.
[47]Prescott B, McLean G F. Line-based correction of radial lens distortiont[J]. Graphical Model and Image Processing,1997,59(1):39-47.
[48]Loomis J G, Palmer J D. Performance Development of a Real-time Vision System[D]. USA:Cornell University,2003.
[49]赵卓,侯义斌,刘静.视觉增强环境中图像校正算法的研究[J].计算机工程与应用,2004,90-92.
[50]徐丽俊,徐小勇基于FPGA的足球机器人数据采集和处理系统仪表技术2008年第3期
[51]郭洪利陈安基于FPGA的高速图像采集／处理卡《福建电脑》2007年2期起止页码：146-146,152
[52]Roger Y. Tsai, A versatile Camera Calibration Technique for High-Accuracy 3D Machine Vision Metrology Using Off-the-Shelf TV Cameras and Lenses, IEEE Journal of Robotics and Automation, Vol. RA-3, No.4,pp.323-344,1987
[53]Tsai, R. Y. A versatile camera calibration technique for high accuracy 3D machine vision using off-the-shell TV cameras and lenses. [J] IEEE Journal of robotics and Automation, Vol. RA-3, NO.4, August 1987:323-344.
[2]RoboCup Organization Official Home page http://www.robocup.org[Z]
[3]The F180 Technical Committee. http://small-size.informatik. uni-bremen. de/rules:main This is an electronic document. Data retrieved:July 20,2009. Laws of the F180 League 2009, Law 2-The Ball and Law 4-The Robotic Equipment,2009, pages 5 and 7.
[4]刘作时,林桂娟,张海英.足球机器人的视觉系统的研制[J].传感器技术,2004,23(3)：38-40.
[5]郭成果.在不稳定光照条件下足球机器人视觉系统的设计[J].机器人：2006-02-12.
[6]谭然.光强自适应的足球机器人视觉系统的设计研究[D].西北工业大学,2007：13-16
[7]付昱强.基于FPGA的图像处理算法的研究与硬件设计,南昌大学硕士论文,2006： 58页
[8]董兵.FPGA图像处理板设计,哈尔滨工程大学硕士论文,2007：4页
[9]Z.-K. Huang and D.-H. Liu, Segmentation of color image using EM algorithm in HSV color space, [J] in Proceedings of IEEE International Conference on Information Acquisition, Jul.2007, pp.316-319.
[10]许志倩,张作龙,朱万胜.足球机器人视觉系统目标识别算法的改进,流体传动与控制,总第22期,第2页。
[11]陈景航.MiroSot足球机器人视觉子系统的改进算法研究.广东工业大学硕士学位论文.2006：20-24页.
[12]Studio encoding parameters of digital television for standard 4:3 and wide-screen 16:9 aspect ratios. http://inst.eecs.berkeley.edu/-cs150/Documents/ITU601. PDF This is an electronic document. Data retrieved:July 25,2009.
[13]Interface for digital component video signals in 525-line and 625-line television operating at the 4:2:2 level of recommendation ITU-R BT.601 (Part A). http://inst.eecs.berkeley.edu/-cs150/Documents/ITU656.PDF This is an electronic document. Data retrieved:July 25,2009.
[14]Yang Liu, Smail Menani, and Johan Dams. Botnia Team Description Paper for RoboCup 2007 Small Size League [R] 2007
[15]Yanowitz S D, Bruckstein A M. A new method for image segmentation. Computer Vision, Graphics, and Image Processing,1989,46:82-95
[16]周鲜成.图像分割方法及其应用研究综述[J]信息技术2007.12
[17]Bruce J, Balch T, Veloso M. Fast and Inexpensive Color Image Segmentation for Interactive Robots [A]. Proceedings of the 2000 IEEE/RSJ International Conference on Intelligent Robots and Systems[C]. USA:IEEE,2000.2061-2066
[18]James Bruce. Realtime machine vision perception and prediction. Undergraduate thesis, Camegie Mellon University May 2000
[19]Burce J , Veloso M. Fast and accurate vision-based Pattern detection and Identification[A]. Proceedings of the 2003 IEEE International Conference on Robotics and Automation[C].USA:IEEE,2003.1277-1282.
[20]William O. McCalister, Chih-Cheng Hung. Image Segmentation Using Run-Length Coding Technique[C].SCIA 2003,LNCS 2749,343-348
[21]严学强,叶秀清,基于量化图像直方图的最大熵阈值处理算法,[J]：模式识别与人工智能,1998.3
[22]周德龙,潘泉,最大熵阈值处理算法,[J]：软件学报,2001.9
[23]Dah-Jye Lee, James Archibald, Xiaoqian Xu and Pengcheng Zhan. Using distance transform to solve real-time machine vision inspection problems. SpringerLink.2006.
[24]Maria E. Angelopoulou, Christos-Savvas Bouganis, Peter Y. K. Cheung and George A. Constantinides. FPGA-Based Real-Time Super-Resolution on an Adaptive Image Sensor. SpringerLink.2006.
[25]李铁军ROBOCUP中型组机器人视觉系统的研究与开发山东大学硕士论文2006年11月
[26]Liang M, Sun Z K. A neural network model for image segmentation. ICAI&NN92, Switzerland.
[27]张建岭,刘军,基于小波变换和形态学的图像分割方法,[J]中国人民公安大学学报(自然科学版),2007.02
[28]胡慧之FPGA在图像处理器中的应用《信息技术》2006年31卷8余楠,孙芳,基于改进遗传算法的图像分割,[J]：计算机与数字工程,2007.8
[29]朱志刚,等译.数字图像处理[M].北京：电子工业出版社,1998
[30]曹爱增,石兴广,刘峰,等.机器人足球颜色分类研究[J].济南大学学报(自然科学版),2005,19(3)：239-241.
[31]Kenneth R. Castleman著,朱志刚等译.数字图象处理[M].北京：电子工业出版社.2002.
[32]Rafael C. Gonzalez, Richard E. Woods,阮秋琦阮宇智等译.数字图象处理(第二版)[M].北京：电子工业出版社.2003.
[33]Yasuda, G. Bin Ge, Color based object recognition, localization, and movement control for multiple wheeled mobile robots, [J]Industrial Electronics Society, 2004. IECON 2004.30th Annual Conference of IEEE
[34]Parag H. Batavia, Sanjiv Singh. Obstacle Detection Using Adaptive Color Segmentation and Color Stereo Homography. Robotics and Automation,2001. Proceedings 2001 ICRA. IEEE International Conference on
[35]黄晶,赵臣,周明明.基于快速彩色空间变换的足球机器人目标搜索[J].哈尔滨工业大学学报.2003.Vol 35 No.9：1036-1039.
[36]杜啸晓,杨新,施鹏飞.机器人足球视党系统的软件研究[J]数据采集与处理.2001第16卷第2期：160-164.
[37]赵小松,张宏伟,张国雄,李真.摄像机标定技术的研究[J].机械工程学报,2002,38(3)：149-151
[38]吴文琪,孙增圻.机器视觉中的摄像机定标方法综述[J]计算机应用研究2004.2
[40]韩娟娟邓文怡娄小平 基于FPGA的图像增强处理系统的设计与实现《微计算机信息》2007年26期起止页码：229-230,122
[41]Faig,W., Calibration of Close-Range Photogrammetry System:Mathematical Formulation, [J], No.12, December 1975,pp.1479-1486.
[42]Wilfried Kubinger, Franz Rinnerthaler, Christoph Sulzbachner, Josef Langer and Martin Humenberger. An Embedded Vision Sensor for Robot Soccer. SpringerLink.2007.
[43]Z. Zhang and C. Loop, Estimating the Fundamental Matrix by Transforming Image Points in Projective Space, Computer Vision and Image Understanding, Vol.82, No.2, pages 174-180,2001
[44]Zhengyou Zhang, Flexible Camera Calibration By Viewing a Plane From Unknown Orientations [J]. Microsoft Research, One Microsoft Way, Redmond, WA 98052-6399, USA. Computer Vision, IEEE International Conference on,vol.1.
[45]杨泽辉孙运强姚爱琴基于FPGA的图像增强压缩系统的设计 《黑龙江科技信息》2007年10期起止页码：85-85
[46]Frederic Devernay, Olivier Faugeras. Straight lines have to be straight Automatic calibration and removal of distortion form scenes of structured environments[J]. Machine Vision and Applications,2001,13:14-24.
[47]Prescott B, McLean G F. Line-based correction of radial lens distortiont[J]. Graphical Model and Image Processing,1997,59(1):39-47.
[48]Loomis J G, Palmer J D. Performance Development of a Real-time Vision System[D]. USA:Cornell University,2003.
[49]赵卓,侯义斌,刘静.视觉增强环境中图像校正算法的研究[J].计算机工程与应用,2004,90-92.
[50]徐丽俊,徐小勇基于FPGA的足球机器人数据采集和处理系统仪表技术2008年第3期
[51]郭洪利陈安基于FPGA的高速图像采集／处理卡《福建电脑》2007年2期起止页码：146-146,152
[52]Roger Y. Tsai, A versatile Camera Calibration Technique for High-Accuracy 3D Machine Vision Metrology Using Off-the-Shelf TV Cameras and Lenses, IEEE Journal of Robotics and Automation, Vol. RA-3, No.4,pp.323-344,1987
[53]Tsai, R. Y. A versatile camera calibration technique for high accuracy 3D machine vision using off-the-shell TV cameras and lenses. [J] IEEE Journal of robotics and Automation, Vol. RA-3, NO.4, August 1987:323-344.