基于DBSCAN的风电叶片音频分类研究

  • 打印
  • 收藏
收藏成功


打开文本图片集

摘  要:利用某风场实际采集的风电叶片音频数据,提取时域与频域特征,利用DBSCAN聚类方法对音频进行分类,可有效区分出机组叶片是否处于运转工况,并且可提取出含噪的音频,有助于针对性的音频去噪工作。

关键词:音频;时域;频域;DBSCAN;分类

中图分类号:TP391 文献标志码:A          文章编号:2095-2945(2022)04-0023-03

Abstract: Using the actual wind turbine blade audio data collected in a wind field, the time domain and frequency domain features are extracted, and the DBSCAN clustering method is used to classify the audio, which can effectively distinguish whether the turbine blade is in operating condition or not, and the noisy audio can be extracted, which can contribute to the targeted audio denoising work.

Keywords: audio; time domain; frequency domain; DBSCAN; classification

风能作为一种环保、绿色、可再生的清洁能源,在全球节能减排进程中起到了越来越重要的作用。(剩余3034字)

目录
monitor