Real-Time Image Processing for Marine Applications

This special issue of the CFPS Journal of Real-Time Image Processing highlights the latest advancements in real-time image processing for marine and underwater applications, including visibility improvement, object detection, and robust perception techniques.

Explore the latest advancements in real-time image processing for marine and underwater applications, in this special issue of the CFP Journal of Real-Time Image Processing. The issue delves with a focus on real-time underwater image enhancement and restoration, techniques, such as visibility improvement, color correction, and noise removal. It also highlights the use of CNN-based enhancement models and Retinex algorithms for real-time object detection, recognition and tracking in underwater environments. like real-time identification of underwater objects including marine species, coral structures and debris using YOLO, Faster R-CNN and lightweight neural networks. Additionally, it covers real-time vision-based navigation and simultaneous localization and mapping (SLAM) for autonomous underwater vehicles (AUVs) and real-time semantic segmentation and scene understanding for marine and underwater environments using techniques like U-Net and DeepLab for coral reef mapping and sea-floor classification. Finally, it discusses real-time multi-modal data fusion integrating visual data with sonar, LiDAR and other sensing modalities for robust underwater perception and lightweight and hardware-accelerated algorithms for deployment on edge devices in resource-constrained underwater environments.

Tags: real-time image processing, marine applications, underwater environments, autonomous underwater vehicles, semantic segmentation, multi-modal data fusion