Comparative analysis and performance evaluation of underwater cable detection and tracking techniques: A comprehensive survey

Abstract

This survey provides a comprehensive review of underwater cable detection and tracking literature, identifying key problem types and highlighting unique underwater challenges. It emphasizes the critical role of underwater cable detection in global communications and energy infrastructures, addressing complexities like low visibility and variable sea conditions. The analysis compares the efficacy of various models, particularly deep learning approaches like CNNs and Transformers, in adapting to underwater imagery challenges. A new roadmap for efficient cable detection and tracking systems is proposed, focusing on multimodal data integration and nonoptical detection methods. Importantly, the study includes performance evaluations of state-of-the-art models on custom underwater datasets, offering practical insights. The survey's findings are validated through an implementation of an underwater object-tracking model incorporating effective algorithms from the literature.