A Quick Classification for Area Environmental Audio Data Based on Local Search Tree

In area environment, various forms of life can exhale different kinds of audio signals, and these audio signals are closely related to a variety of organisms living environment and human activities. For analyzing these audio data automatically, local search data structure is proposed in this paper, that it is local search tree (LS-Tree), to classify audio data in the natural environment rapidly. Firstly, a technology for classification is brought forward in this paper using effective segment length of audio data (ESLOAD), frequency component of maximum harmonic weight (FCOMHW) and first order difference Mel-frequency cepstral coefficients matrix (D-MFCCM) in a segment. Secondly, a LS-Tree is pro- posed in this paper. ESLOAD and FCOMHW in segment are chose as main and second key for the LS-Tree respectively. Classification parameters are classified and managed through the LS-tree. Finally, through experiment of nine categories totally 107 normal audio data, it is indicated that the technique is effective to classify many kinds of audio data in the natural environment.