波形模式匹配的一种加速算法

A Fast Algorithm for Waveform Pattern Matching

  • 摘要: 本文提出了用于某些波形识别模板匹配的一种模式识别与阈值加速算法.这种方法的基本思想是将波形的幅度期望值和波形的上升、下降、水平的波动特征构成筛选阈值和特征向量,对已知的波形集合进行筛选,仅对满足阈值要求并且与特征向量完全相同的波形才进行模板匹配的相关计算.对于有一定特征的波形集合信息压缩率达到90~98%.提高计算速度10~30倍.

     

    Abstract: An algorithm for pattern recognition and the calculation of the threshold valuehas deen proposed to speed up the pattern matching of certain waveforms.The main idea is that from the expected amplitude and the vertical and horizontal fluctuation of the waveform a characteristic vector is derived and the thres hold value calculated.For certain a set of wave forms the matching operation is carried outonly for those waveforms that have the same vector as the pattern,and are within thethreshold value.For some wavefoms,the information compression ratio reaches 90~98%,the speed of the computation is increased by a factor of 10—30.

     

/

返回文章
返回