An Approach to Detect Dolphin Whistle from Underwater Sounds

This paper handles the issue of identifying the whistle of a dolphin in the presence of noise. When a recorded underwater signal containing other sounds is given to a system, it detects and identifies the presence of dolphin whistle. Analyzing the extracted underwater sounds gives meaningful information about the environment. Also, the presence of a particular underwater species and their behavior can be identified from the sounds emitted by them. Dolphins make distinctive calls. This can be analyzed in the post processing stage. Features like Mel-frequency cepstral coefficients are extracted for the detection of dolphin whistle. Here, the classifier Gaussian Mixture Model is used to classify the recordings based on features to identify the presence of dolphin whistle. A complete simulation is done on the recorded underwater signals to detect the presence of dolphin. Index Terms — Underwater, Detection, GMM, Dolphin Whistle ——————————  ——————————