Firstly, phonemes should be extracted on the time domain, the frequency domain and energy then the combination character of each sample and map figure were calculated.
首先对语音音素进行时域、频域、能量值的提取，然后通过计算得到每个采样点的组合特征值，描绘成图像。
Having good time-frequency localization character, and correctly identifying singularity point in fault signal, Wavelet is the main analysis method in this paper.
小波变换具有良好的时频局部性，具有变焦距的特点，对故障信号中的奇异点能够准确识别，是论文中应用的主要分析方法。
Fourth, establish a "record of weak character investigation", fill in the content, including time, frequency, weak character state, the investigation of people, and disposal method.
四是建立“薄弱人物排查记录”，填写内容包括时间、班次、薄弱人物状态、排查人、处理方式等。
应用推荐