摘要
为了提高禽类养殖业的自动化水平,实现雏鸡性别的自动识别,提出了一种基于音频分析技术的雏鸡性别识别方法。采用自相关法提取雏鸡鸣叫声的基音频率作为信号的特征参数。通过隐·马尔可夫模型对基音频率进行识别分类,实现性别识别。实验结果表明,当采用多数投票的识别策略时,识别率超过人工识别。
In order to improve the automation level of the poultry breeding industry,to realize the automatic recognition of sex of chicken,put a method of chicken sex identification based on the technology of audio analysis.Use the autocorrelation pitch frequency extraction chick calls as the signal characteristic parameter.Through the hidden Markov model,to classify the Pitch frequency recognition and achieve gender recognition.The experimental show that,when use the majority voting recognition strategy, the rate of recognition exceed artificial recognition.
出处
《电子测试》
2014年第1X期107-108,共2页
Electronic Test
