Answers are grounded in this paper and cite their sources.
Reliable underwater object detection is still a very difficult task because of the environmental distortions, including turbidity, scattering of light, and color loss. In order to overcome these challenges, this research introduces an IoT-environment adaptation-boosted Context-Aware YOLO Transformer Fusion Model. The model is a combination of the real-time localization ability of YOLO with the global dependency modeling ability of a Transformer encoder that gives a hybrid detection framework robust against dynamic underwater environments. IoT sensors constantly record contextual information, such as temperature, turbidity, salinity and light, which are combined with visual features to dynamically adjust the Transformer's attention weights. The cross-modal feedback mechanism allows the system to sensitively adjust detection thresholds and feature representations as a function of changing water clarity and lighting conditions. It was trained on an Underwater Object Dataset of 638 annotated images in seven marine categories. On experimental results, detection accuracy, recall, and mean Average Precision (mAP) are demonstrated in comparison with the state-of-the-art baselines of YOLOv7 and Swin-Transformer. It achieves the 97.4% detection accuracy and a 0.08 loss without overfitting. The suggested hybrid system offers a flexible, context-aware, system for autonomous underwater surveillance, marine ecology assessment, and intelligent robotic exploration.
No takes yet. Share an insight, caveat, or question.
Neelamegam et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: