音频的自动分类,尤其是语音和音乐的分类,是提取音频结构和内容语义的重要手段之一,它在基于内容的音频检索、视频的检索和摘要以及语音文档检索等领域都有重大的应用价值.由于隐马尔可夫模型能够很好地刻画音频信号的时间统计特性,因此,提出一种基于隐马尔可夫模型的音频分类算法,用于语音、音乐以及它们的混合声音的分类.实验结果表明,隐马尔可夫模型的音频分类性能较好,最优分类精度达到90.28%. The automatic classification of audio, especially the classification of speech and music, is one of the important means of extracting audio structure and content semantics. It has great applications in content-based audio retrieval, video retrieval and summarization and speech document retrieval Because Hidden Markov Model can characterize the time statistics of audio signal well, this paper proposes an audio classification algorithm based on Hidden Markov Model for the classification of speech, music and their mixed sounds. The results show that Hidden Markov Model has better audio classification performance and the best classification accuracy is 90.28%.