论文部分内容阅读
在Miallat塔式分解算法基础上,提出流式Mallat塔式分解算法,该算法克服了Mallat塔式分解由于采用逐级二次抽样造成定位检测误差的缺点.根据数据流信号的特点,选择适当的共轭镜像滤波器对流式分解的每一级都能准确跟踪数据流中的每一点.计算机仿真实验对比结果表明,流式Mallat分解算法用于事件检测、局部分析等方面比原Mallat分解算法更具优越性.“,”Flowing Mallat’s pyrarnidal decomposition algorithm is presented based on the orginal Mallat’s pyramidal decomposition algorithm. It overcomes the shortcoming of location bias caused from subsample by two (drop every other sample) in each level of the original Mallat’s pyramidal decomposition algorithmss,Choosing proper quadrature mirror filters according to the characteristics of signals, this new decomposition algorithm can lo cate accurately every point of data flow in each level of decompostion. Computer simulations results show that our flowing Mallat’s decomposition algorithm is much better for event detections, local analyis and so on than the original Mallat algorithm.