Underwater Passive Target Recognition Based on Self-Supervised Contrastive Learning

In recent years, deep learning has garnered widespread attention in the field of underwater passive target recognition. Existing recognition methods mostly rely on learning from a number of labeled samples, leading to performance degradation when dealing with the small sample characteristics of scarce high-quality underwater acoustic data. Additionally, the utilization of the large amounts of unlabeled data in underwater acoustic datasets is relatively low. To address these issues, this paper proposes using self-supervised learning to perform feature representation learning on unlabeled data, followed by supervised fine-tuning of the pretrained feature extractor using a small amount of labeled data. Validation on real-world datasets demonstrates the effectiveness of this approach.