快速飞行的乒乓球在图像采集过程中会产生拖影, 影响甚至严重干扰对球体位置的准确识别和跟踪。对飞行乒乓球的拖影图像特点及实战型乒乓机器人的实际需求进行分析, 在此基础上提出CS-BS-EF方法, 将色彩分割、背景减除、椭圆拟合进行综合集成、优势互补, 用以检测飞行球体的拖影范围并求取球心位置。实验结果表明该方法能够有效克服拖影影响, 比较准确地识别和跟踪高速飞行的球体, 且具有良好的抗干扰性、实时性。
Fast flying ping-pong ball usually generates motion blur in image capturing process, which affects and even interferes in the process of accurately recognizing and tracking their positions.In this paper, motion blur features of flying ping-pong ball, as well as the actual demands by an adversarial ping-pong playing robot, are analyzed.Then we propose an integrated method called CS-BS-EF, which stands for a comprehensive integration and optimization in methods of color-segmentation, background-subtraction and ellipse-fitting, which overcomes the drawbacks of each one.Now, CS-BS-EF is used for detecting smear scope and calculating and tracking flying ping-pong ball.The experimental results show that CS-BS-EF is capable of surmounting the effects of motion blur, and of recognizing and tracking high-speed flying ping-pong ball with relative accuracy, anti-interfering, and instantaneity.
[1]Acosta L, Rodrigo J J, Mendez J A, et al.Ping-pong player prototype[J].IEEE Robotics & Automation Magazine, 2003, 10(4):44-52.
[2]Rusdorf S, Brunnett G, Lorenz M, et al.Real-time interaction with a humanoid avatar in an immersive table tennis simulation[J].IEEE Transactions on Visualization and Computer Graphics, 2007, 13(1):15-25.
[3]Miyazaki F, Takeuchi M, Matsushima M, et al.Realization of the table tennis task based on virtual targets[C].Washington DC:Proceedings of the 2002 IEEE International Conference on Robotics and Automation, 2002:3844-3849.
[4]彭博, 洪永潮, 杜森森, 等.乒乓球机器人击打点的预测方法[J].江南大学学报(自然科学版), 2007, 6(4):433-437.
[5]Wilson A.Vision-based robot/human ping-pong[J].Vision Systems Design, 2008, 13(2):21-22.
[6]郑魁敬, 崔培.乒乓球机器人的研究与发展[J].机床与液压, 2009, 37(8):238-241.
[7]杨绍武.基于双目视觉的乒乓球识别与跟踪问题研究[D].北京:国防科技大学, 2009.
[8]张远辉.基于实时视觉的乒乓球机器人标定和轨迹跟踪技术研究[D].杭州:浙江大学, 2009.
[9]Zhang Yu-ye, Wang Xue-wei, Wang Chun-xin.Restoration of space-variant blurred image based on motion-blurred target segmentation[J].Journal of Systems Engineering and Electronics, 2010, 21(2):191-196.
[10]张进, 陈慧蓉, 荣钢.基于波方程的运动模糊图像恢复[J].清华大学学报(自然科学版), 2005, 45(7):1002-1004, 1008.
[11]程姝, 赵志刚, 吕慧显, 等.顺序结构的运动模糊图像复原技术综述[J].计算机应用, 2013, 33(S1):161-165, 185.
[12]刘微.运动模糊图像恢复算法的研究与实现[D].长春:中国科学院长春光学精密机械与物理研究所, 2006.
[13]林开颜, 吴军辉, 徐立鸿.彩色图像分割方法综述[J].中国图象图形学报, 2005, 10(1):1-10.
[14]李偲, 栾新, 宋大雷.基于彩色图像分割的机器人的目标识别[J].微计算机信息, 2008, 24(6-2):358-360.
[15]Tan K S, Isa N A M.Color image segmentation using histogram thresholding-Fuzzy C-means hybrid approach[J].Pattern Recognition, 2011(44):1-15.
[16]Dana E I, Paul F W.Image segmentation based on the integration of colour-texture descriptors-A review[J].2011(44):2479-2501.
[17]Zhou Hong, Chen Yi-ru, Feng Rong.A novel background subtraction method based on color invariants[J].Computer Vision and Image Understanding, 2013, 117(11):1589-1597.
[18]Yang Jin-bin, Shi Min, Yi Qing-ming.A new method for motion target detection by background subtraction and update[J].Physics Procedia, 2012, 33:1768-1775.
[19]李刚, 邱尚斌, 林凌, 等.基于背景差法和帧间差法的运动目标检测方法[J].仪器仪表学报, 2006, 27(8):962-964.
[20]韩亚伟, 张有志, 李庆涛, 等.帧差累计和减背景相结合的运动对象分割方法[J].计算机工程与应用, 2011, 47(14):164-166.
[21]闫蓓, 王斌, 李媛.基于最小二乘法的椭圆拟合改进算法[J].北京航空航天大学学报, 2008, 34(3):295-298.