基于FPGA的全自动定位捡乒乓球机器人系统设计  

Design of automatic positioning and picking up ping⁃pong ball robot system based on FPGA

作　　者：孟群康 杨凯宁 李井源 江先阳[1,2] MENG Qunkang;YANG Kaining;LI Jingyuan;JIANG Xianyang(School of Physical Science and Technology,Wuhan University,Wuhan 430072,China;National Demonstration Center for Experimental Physics and Education,Wuhan University,Wuhan 430072,China)

机构地区：[1]武汉大学物理科学与技术学院,湖北武汉430072 [2]武汉大学物理国家级实验教学示范中心,湖北武汉430072

出　　处：《电子设计工程》2023年第12期28-32,37,共6页Electronic Design Engineering

摘　　要：对现有电子辅助的运动器材行业进行分析可以发现,目前市场上应用于场馆训练的大量散落乒乓球主要依靠人力或者简单的工具拾捡,而已有的捡乒乓球机器人很多智能化程度不高、识别不够精确。为此,提出一种基于FPGA(Field Programmable Gate Array)的全自动定位捡乒乓球机器人的系统设计方案,利用摄像头采集机器人前方的图像数据实现乒乓球定位,利用超声波传感器实现机器人的自动避障,利用机械臂完成抓取乒乓球动作。图像识别乒乓球的部分创新性地将Hough变换算法硬件化,有效提高识别率和捡球速度。测试结果表明,机器人在定制测试条件下可以达到97.8%的识别准确率,具有自动寻球、自动捡球、自动避障以及手机蓝牙便捷开关机的功能。系统可以快速方便应用到乒乓球球馆中,市场潜力很大。Based on analysis of the existing electronically assisted sports equipment industry,it can be found that a large number of scattered ping⁃pong balls in the stadium are mainly picked up by manpower or simple tools,and most of the existing pick⁃up ball robots do not have high level of intelligence or accuracy.In order to solve this problem,a system designed for automatically locating and grabbing ping⁃pong balls based on FPGA(Field Programmable Gate Array)is proposed,which utilizes camera to detect the front image of robot to find the location of ping⁃pong balls,applies ultrasonic sensors to avoid obstacles automatically,and takes advantage of robotic arms to carry out the action of grasping ping⁃pong balls.The system innovatively implements Hough transform in Hardware Description Language to improve the ball recognition accuracy and picking up speed.Experimental results show that the robot can achieve 97.8%recognition accuracy under custom test conditions with features of automatically finding and picking ping⁃pong balls,avoiding obstacles,and users can control it on mobile phone by Bluetooth conveniently.The proposed system can be applied in stadium quickly and has great market potential.

关 键 词：FPGA 可重构机器人 自动避障 图像处理 SOBEL边缘检测 HOUGH变换 

分 类 号：TN431.2[电子电信—微电子学与固体电子学]