基于电容的手势识别系统

传统手势识别装置一般依赖于复杂计算设备,在简单场合配置手势识别装置成本过高。针对这一问题,提出了基于电容数字转换器的手势识别系统。具体实施过程中,电容数字转换器采集静态手势特征,滑动窗口滤波加工特征数值,再结合K邻近分类算法判别静态手势。该系统具有对硬件要求低、功耗小、对环境光强不敏感等优点。经过50次试验,系统对于常用的数字零到五手势识别正确率为98%,判别时间小于0.6 s。

Design of the Capacitor Based Gesture Recognition System

Commonly used gesture recognition system is too costly to be applied in scenarios with low computing resources.A capacitor-based gesture recognition system is proposed to solve this problem.Raw gesture data is captured by the capacitance converter and then processed with Sliding-Windows filter.Final gesture mark is given by KNN(K Nearest Neighbor)classification algorithm.The proposed method features low hardware requirement,low power consumption and non-sensitivity to light.Experimental results showed that the system can achieve gesture recognition rate up to 98%with time consumption less than 0.6s for each recognition process.