论文部分内容阅读
声音分析技术已成为研究动物行为、动物福利的一种重要工具,蛋鸡声音可用来评价其健康或福利状况,然而规模化蛋鸡舍中存在较多风机噪声等声源干扰,这对准确地识别蛋鸡发声有很大影响。本文以海兰褐蛋鸡为例,预处理获取的声音信号,减少风机噪声的干扰。利用数字化声音采集平台采集不同类型的蛋鸡发声和风机噪声音频,采用LabVIEW软件进行声音信号处理并分析蛋鸡声音和风机噪声的时频特征。同时,对比分析不同去噪方法(IIR滤波器去噪、小波阈值去噪和改进谱减法去噪)在去除风机噪声方面的效果。结果表明,在信噪比为-8~20 dB声音环境情况下,改进谱减法均方根误差最小(0.03~0.38),算法运行耗时最短(6~7 ms),在实际应用中去噪效果较好。该研究可为规模化蛋鸡舍中风机噪声环境下的蛋鸡声音信号处理和分析提供参考。
Sound analysis techniques have become an important tool for studying animal behavior and animal welfare, and the sound of laying hens can be used to evaluate their health or welfare status. However, there are more noise sources such as fan noise in large scale egg henhouse, which is accurate Recognizing the hen’s voice has a big impact. In this paper, Hy-line hen was taken as an example to preprocess the acquired voice signal to reduce the interference of fan noise. Using digital sound collection platform to collect different types of hen and fan noise audio, using LabVIEW software for sound signal processing and analysis of chicken noise and time-frequency noise characteristics. At the same time, the effects of different denoising methods (IIR filter denoising, wavelet threshold denoising and improved spectral subtraction denoising) were compared and analyzed in terms of removing fan noise. The results show that the RMSE of the improved spectral subtraction is the smallest (0.03-0.38) when the signal to noise ratio is -8-20 dB, and the shortest operation time (6-7 ms) is required in the practical applications The effect is better. This study may provide a reference for the processing and analysis of laying hen’s sound signals in a large-scale laying henhouse fan noise environment.