Recognition of Fruit Fly Wings Vibration Sound Based on HMM

This paper classifies and recognizes three different strains of fruit fly by their wings vibration sounds. It uses Mel-Frequency Cepstrum Coefficient to extract features of fruit fly wings vibration sound, then applies Hidden Markov Model to establish models of three different strains of fruit fly wings vibration sound and recognize three strains of fruit fly. Experiment shows that the recognition rates of the three strains of fruit fly reach over 70%, and the result is satisfying. Thus, the method used by the paper is feasible and effective, and provides a new basis for fruit fly´s research.