摘要
介绍了谱消除法、小波变换和数学形态学滤波三种信号处理方法进行消噪的基本原理,并分别采用这三种方法对微机械陀螺的静态漂移信号和动态仿真信号进行了去噪声处理,对处理前后的信号误差大小及其分布情况以及处理过程的运算量和实时性进行了比较。结果表明,对于静态漂移信号的处理,三种方法的滤波效果大致相当。对于动态信号,数学形态学滤波器的滤波效果比其它两种滤波器的滤波效果要好,并且,数学形态学滤波器具有运算量小,实时性较好的优点,十分易于硬件实现。
The principles of three signal processing methods of clearing the out-of-band spectrum, fast wavelet transformation and the mathematical morphology filtering for de-noise were introduced. These three methods were also used to de-noise the static drift signal and the dynamic simulation signal from the gyroscope. Comparisons of the noise error and the error distribution were made among the different methods. The simulation result showed that the static drift signal filtering effects of the three filtering methods were almost the same. But to the dynamics measurement, the mathematical morphology had the best filtering effect. And the virtues of the less procession burden and the less real-time processing of the mathematical morphology filtering made it very easy for its hardware implementation.
出处
《高技术通讯》
CAS
CSCD
北大核心
2007年第3期274-278,共5页
Chinese High Technology Letters
基金
国家863计划（2002AA812038）资助项目.
关键词
小波变换
形态学滤波
微机械陀螺
wavelet transformation, morphology filtering, micro-machined gyroscope
