Semantic segmentation-based whistle extraction of Indo-Pacific Bottlenose Dolphin residing at the coast of Jeju island

•Underwater acoustic is widely used to monitor dolphins.•Noise at coastal habitats increases the difficulty of dolphin vocalization detection.•Semantic segmentation classifies the dolphin whistle with the appearance of a shrimp snapping sound.•Morphology analysis extracts actual dolphin whistle from tangled contours.•The method introduced detects the dolphin whistle with a Precision of 0.96, Accuracy of 0.9, and an F-score of 0.86.