首页|基于人体关键点识别的机器人舞姿模仿系统研究

基于人体关键点识别的机器人舞姿模仿系统研究

扫码查看
随着机器人技术的持续进步,使其能模仿人的舞蹈动作呈现出重要的研究价值.此次研究致力于开发基于人体关键点识别的机器人舞姿模仿系统.研究采用了基于神经网络的方法,首先构建二维关键点识别模型,再利用回归模型从二维坐标中推断三维坐标,并结合逆运动学技术将其转化为机器人的动作指令.研究数据表明,对于头部、肩部、肘部和髋关节的识别准确率分别达到95.8%、94.5%、87.39%和87.3%.而在三维坐标预测中,平均偏差为43.91 mm.尽管此误差在某些领域仍然存在挑战,但由于机器人尺寸较小,其实际影响相对较小.结果证明,研究提出的方法在机器人舞蹈模仿上具有很好的效果和应用潜力.
Research on robot dance posture imitation system based on human key point recognition
With the continuous progress of robot technology,it has shown important research value to make it imitate human dance movements.This research is devoted to developing a robot dance posture imitation system based on human key point recogni-tion.The research adopts the method based on neural network.Firstly,the two-dimensional key point recognition model is construc-ted,and then the regression model is used to infer the three-dimensional coordinates from the two-dimensional coordinates,and the inverse kinematics technology is combined to transform the action instructions of the robot.The accuracy of head,shoulder,elbow and hip joints was 95.8%,94.5%,87.39%and 87.3%,respectively.In the three-dimensional coordinate prediction,the average deviation is 43.91mm.Although this error remains a challenge in some areas,its practical impact is relatively small due to the small size of the robot.The results show that the proposed method has good effect and application potential in robot dance imitation.

human bodykey pointsrobotsdancing posturerecognize

丁玲、赵昆

展开 >

陕西工业职业技术学院,陕西咸阳 712000

人体 关键点 机器人 舞姿 识别

中国机械政研会机械职业教育思想政治工作研究分会

SZ22B058

2024

自动化与仪器仪表
重庆工业自动化仪表研究所,重庆市自动化与仪器仪表学会

自动化与仪器仪表

CSTPCD
影响因子:0.327
ISSN:1001-9227
年,卷(期):2024.(6)
  • 13