An automatic identification algorithm of Yangtze finless porpoise

This paper describes the time-frequency features of the Yangtze finless porpoise acoustic signals firstly. Then an automatic identification algorithm for the Yangtze finless porpoise is presented based on Hilbert Huang Transform and BP artificial neural network. The algorithm includes two steps: feature extraction and signal identification. In the first step of the algorithm, the algorithm extracts a 10-Dimension signal feature vector based on Hilbert Huang transform, Hilbert marginal spectrum and Fourier transform. In the identification step, the BP artificial neural network is trained by using the extracted vector as input. Some experimental acoustic data files of finless porpoise are used to test the validity of the automatic identification algorithm. 238 finless porpoise acoustic signals are detected. The correct identification probability of the algorithm proposed in this paper reaches 93%, according to the human observation on the time-frequency spectrum. Because the Yangtze finless porpoise is one of the most endangered mammals in the world, so the presented method has great practical significance for surveying and protecting the Yangtze finless porpoise in the wild.