An automatic method to detect the presence of elephant

In this paper an approach has been discussed to automatically detect the presence of elephant by the detection of a low frequency sound produced by elephant called rumble. A detection system has been developed which uses features of rumbles as input to recognize the rumbles. Feature extraction techniques have been used to extract the features of rumble which includes the greenwood function cepstral coefficients (GFCC) and first three formant frequencies. The GFCC features have been extracted in a procedure similar to mel frequency cepstral coefficients (MFCC) extraction while the formant frequencies have been extracted using linear predictive coding (LPC). A robust feature vector has been created by cascading the formant frequencies to the GFCC features. The detection system has been developed using feed forward neural network which is trained using backpropagation algorithm.