Identification of whale populations emerges as a crucial initiative in conservation efforts. While most existing methods are invasive and costly, there is a growing need for more practical, non-invasive approaches to facilitate whale identification. This study explores the use of acoustic signals produced by whales as a minimally invasive, data-driven alternative, combined with machine learning techniques. It is important to note that the objective here is not to identify specific whale species, but rather to distinguish whale vocalizations from other anthropogenic oceanic acoustics. As whales, being marine mammals, emit patterned sounds during mating and foraging activities, their vocalizations present an opportunity for systematic identification. Therefore, this study adopts Music Information Retrieval (MIR) techniques to extract relevant audio features from waveforms of oceanographic acoustic data. These features were preprocessed and used to train and validate a binary classification model using the XGBoost algorithm, which initially achieved an overall accuracy of 92%. However, further evaluation through precision, recall, and F1-score revealed that the model performs significantly better in recognizing non-biological sounds than in detecting whale songs. This discrepancy is attributed to the high class imbalance in the dataset, where whale vocalizations were recorded only 297 times, compared to 2,694 instances of anthropogenic sounds. To address this imbalance, data rebalancing techniques such as Random Oversampling, SMOTE, and ADASYN were applied. Among these, Random Oversampling demonstrated the best results, achieving the highest F1-score of 63% for the minority class and improving overall accuracy to 93%. Hence the study presents the XGBoost model, enhanced with Random Oversampling, as a promising tool for the non-invasive identification of whale populations using oceanographic acoustics. This approach underscores the potential of integrating computer science and machine learning with biodiversity conservation, offering a scalable solution for long-term monitoring of Whale populations.