摘要
针对人机交互中对于动态手势识别技术的需求,设计了一种新的平面电容传感阵列对动态手势进行检测。通过COMSOL Multiphysics对传感器阵列结构进行仿真设计,采用卡尔曼滤波算法对手部运动坐标数据进行滤波,并借鉴基于视觉的手势识别中的隐马尔可夫模型(HMM)方法对手部运动轨迹进行识别。该系统对动态手势图形的平均识别率高达95%,可用于简单的人机交互。
Aiming at the requirements of dynamic gesture recognition technology in human-computer interaction,a flat capacitor sensing array is used to simulate the sensor array structure through COMSOL Multiphysics,a new plane capacitor sensing array is designed,and the HMM(Hidden Markov Model)method in visual gesture recognition is used to filter the motion coordinates of the opponent part of Kalman filtering,and then the opponent movement trajectory is identified by using the HMM method.The final implementation of the recognition of simple graphics.The average recognition rate of graphics is as high as 95%.
作者
王子懿
沈三民
佘硕铖
WANG Ziyi;SHEN Sanmin;SHE Shuocheng(Key Laboratory of Instrumentation Science&Dynamic Measurement,Ministry of Education,North University of China,Taiyuan 030051,China)
出处
《测试技术学报》
2023年第1期54-59,共6页
Journal of Test and Measurement Technology
