摘要
将隐马尔可夫模型引入到舰船噪声目标识别中。选择MEL频率倒谱及其差值参数作为模型的观察序列,分别采用Baum-Welch算法和Viterbi算法对模型进行训练和识别。对这一模型进行了计算机模拟,并利用潜艇、鱼雷和水面舰3类目标的海上实录噪声对其识别性能进行了测试,正确识别率为87%。实验表明,用隐马尔可夫模型对舰船辐射噪声进行目标识别可以取得较好的效果。
Hidden Markov Models are introduced to the area of ship-radiated noise recognition.Mel-Frequency Cepstral Coefficients and the differencing parameters are used as characteristics,and BaumWelch algorithm and Viterbi algorithm are used as training and recognizing method respectively.The model is cyber-simulated and three types of ship noise are used to analyze its recognizing ability.The recognizing rate is 87%.The results show that HMMs are feasible to ship-radiated noise recognition.
出处
《青岛大学学报（自然科学版）》
CAS
2006年第2期34-37,共4页
Journal of Qingdao University(Natural Science Edition)
关键词
舰船
辐射噪声
识别
隐马尔可夫模型
ship
radiated noise
recognition
Hidden Markov Models
