基于数据手套的手势识别系统  

Design of Gesture Recognition System Based on Data Gloves

作　　者：刘雪花 洪琴 何玉麟 林嘉兴 Liu Xuehua;Hong Qin;He Yulin;Lin Jiaxing(Dept.of Computer,South China Institute of Software Engineering,Guangzhou University,Guangzhou 510990,China)

机构地区：[1]广州大学华软软件学院计算机系,广州510990

出　　处：《单片机与嵌入式系统应用》2020年第6期16-19,共4页Microcontrollers & Embedded Systems

基　　金：广州大学华软软件学院质量工程项目(ZYGXK201805);国家级大学生创新创业基金项目(201912618008)。

摘　　要：为了辅助聋哑人与正常人进行正常沟通,设计了一种基于数据手套的手势识别系统。本文首先通过STM32来采集MPU6050加速度传感器和Flex2.2应变片的相关数据,然后利用动态手势滤波进行数据预处理,再采用循环神经网络(RNN)进行手势的识别,最后通过移动终端APP将识别结果以语音和文本的方式输出。该系统一共可以识别20种常用的交流手势,对于聋哑人进行日常交流具有一定的辅助作用。In order to assist deaf-mute people in normal communication with normal people,a gesture recognition system based on data gloves is designed.This design first collects relevant data of MPU6050 acceleration sensor and Flex2.2 strain gauge through STM32,then uses dynamic gesture filtering for data preprocessing,then uses recurrent neural network(RNN)for gesture recognition,and finally through mobile terminal APP Output the recognition result in voice and text.The system can recognize a total of 20 commonly used communication gestures,and it has a certain auxiliary effect on daily communication for deaf people.

关 键 词：数据手套 手势识别 RNN STM32 MPU6050 

分 类 号：TP274[自动化与计算机技术—检测技术与自动化装置]