摘要
在运动训练过程中,采集工作是关系到运动员训练水平的关键。将多传感器的思想应用到信息融合技术上,通过多个传感器和摄像机分别对运动员的关键指标进行采集。前者是利用小波变换实现对人体表面肌电信号的特征提取,然后使用人工神经网络法对关键的特征的进行识别,得到初步人体运动评估结果。后者是对图像视频经过前景提取、特征提取、识别姿态及评估进行处理。提高了人体疲劳评估准确率,从根源上解决了传统系统的采集准确率地下的问题,经过科学指导,提高运动员训练水平,对于信息融合技术具有重要的作用。
In the process of sports training,the collection work is the key to the training level of athletes.The idea of multi-sensor is applied to the information fusion technology.The key indicators of athletes are collected by multiple sensors and cameras.The former uses wavelet transform to extract the features of EMG signals,and then uses artificial neural network to identify the key features,and gets the preliminary results of human motion evaluation.The latter is to process the image video through foreground extraction,feature extraction,attitude recognition and evaluation.It improves the accuracy of human fatigue assessment,solves the problem of the tradi tional system's collection accuracy from the root,improves the training level of athletes through scientific guidance,and plays an im portant role in information fusion technology.
作者
王思
WANG Si(Xi'an Medical University,Xi'an,Shaanxi 710021,China)
出处
《计算技术与自动化》
2020年第3期140-146,共7页
Computing Technology and Automation
关键词
传感器
信息融合技术
小波变换
神经网络模型
证据理论
sensor
information fusion technology
wavelet transform
neural network model
evidence theory
