Efficient Underwater Image Processing for Autonomous Systems

Dear Colleagues,

Accurate underwater scene perception serves as the fundamental prerequisite for comprehensive environmental understanding and effective physical interaction. A thorough understanding of the underwater image formation model is essential for this perception process. Constrained by current technological bottlenecks, underwater vehicles predominantly rely on pre-programmed routines or rudimentary heuristic algorithms for tasks such as underwater navigation, underwater path planning, and underwater obstacle avoidance. Consequently, they generally fall short in autonomously executing operations that demand dexterous manipulation and high-precision positioning. Therefore, developing underwater scene perception technologies with robust performance has emerged as an imperative research frontier in underwater robotics.

Underwater scene perception fundamentally aims to decipher complex scenes while continually integrating novel information to accommodate dynamic, unstructured environments. Despite recent advancements, state-of-the-art perception frameworks remain inadequate for practical robotic deployments, with major technical challenges including rapid processing of underwater images via underwater image enhancement, underwater image restoration, and underwater image segmentation, as well as accurate underwater object detection, underwater object tracking, and underwater object localization.

Therefore, the Editors of this Topic encourage submissions of original research studies and reviews including, but not limited to, the following topics and issues:

• Underwater image enhancement;
• Underwater image restoration;
• Underwater image generation;
• Underwater image segmentation;
• Underwater 3D scene modelling;
• Underwater object detection, recognition, and tracking;
• Visual SLAM and vision-based localization for autonomous navigation;
• Vision-guided path planning and obstacle avoidance;

Domain adaptation and generalization for cross-scene underwater image processing.

Prof. Dr. Mingxing Lin
Prof. Dr. Zhiguang Guan
Topic Editors