高级检索

隐马尔可夫模型在舰船噪声目标识别中的应用

Ship-radiated noise recognition based on CHMM

  • 摘要: 将Mel频率倒谱作为特征参数,实现了基于连续型隐马尔可夫模型的舰船辐射噪声目标识别。对五艘不同类别舰船的辐射噪声建立模型,并进行了识别实验。模型的状态数和高斯混合数分别确定为1和3,五艘舰船的平均识别率达到91.6%。利用两艘不同类型舰船的噪声分析了舰船的工况对识别结果的影响。

     

    Abstract: Ship-radiated noise recognition based on Continuous Hidden Markov Models(CHMM) is achieved.MelFrequency Cepstral Coefficients(MFCC) are used as features.Ship-radiated noises of five different classes are modeled and target recognition experiments are conducted.the number of states and Gaussian mixture is set to be 1 and 3 respectively.The average recognition rate is up to 91.6% for 5 different ships.Noises of two different ships are used to analyze the impact of different operating condi-tions on the recognition rate.

     

/

返回文章
返回