本文构建了一个基于FPGA的实时手势识别平台,并在该平台上实现了一种基于表面肌电(sEMG)信号和加速度(ACC)信号的手势识别算法.具体实现过程中,无线sEMG传感器和无线三轴ACC传感器穿戴于两手前臂实时获取sEMG信号和ACC信号,并以无线方式发送到数据处理模块.数据处理模块充分利用FPGA的并行处理优势,融合ACC和sEMG信息特征,实现了单双手手势的实时识别.经测试,本文所用的手势识别算法移植到FPGA中以后,识别速度明显提高,16个中国手语手势动作达到了95%以上的识别率.%This paper developed a real-time gesture recognition system based on FPGA, and achieved a gesture recognition algorithm based on surface electromyography ( sEMG) and acceleration ( ACC ) signals on the system. Specifically, the wireless sEMG sensors and wireless tri-axial ACC sensors were worn on the arms to collect signals and sent the signals to the data processing module. The processing module took full advantage of FPGA parallel processing,and put sEMG and ACC signal features together to achieve real-time hand gesture recognition. According to the test, the hand gesture recognition algorithm transplanted on FPGA got faster recognition speed than it was run on a personal computer,and the recognition rate for sixteen kinds of Chinese sign gestures reached 95% above.
展开▼