Application of multimodal image processing and object detection algorithms in preventing ship damage to submarine optical cables

Artificial intelligence technology has important application value in the field of submarine cable safety protection. A multi-modal image processing and object detection algorithm is proposed to improve the early warning capability of submarine fiber optic cable against ship damage, in response to the problem of submarine fiber optic cable damage caused by ship activities. This method first uses a residual dense attention network to achieve efficient fusion of multimodal images and enhance the expression of target features. Then, an edge enhanced guided object detection network was constructed, which extracted shallow features of underwater images through an edge extraction module and suppressed noise using a global information extraction module to improve detection accuracy. The results showed that the signal-to-noise ratio of the fused image increased by 12.75%, and the structural similarity increased by 11.48%. Practical application verification shows that the proposed method has a real-time detection accuracy of 96.52% for ship intrusion events, and the false alarm rate has been reduced to 2.61%. This indicates that the raised method can effectively improve image quality, identify ship threats, and achieve accurate warning, providing an efficient technical means for the safety protection of submarine optical cables.