| Underwater Simultaneous Localization and Mapping (SLAM) technology enables underwater robots, to perform self-localization and environmental mapping simultaneously in unknown underwater environments. This technology holds significant importance for fields like oceanographic research and seabed resource exploration. This review paper summarizes the latest research advancements, challenges, and solutions in underwater visual SLAM, as well as the key theoretical aspects of underwater visual SLAM. The complexity of underwater environments, characterized by factors such as light attenuation, scattering, and current disturbances, presents numerous challenges for the research and development of underwater SLAM technology. The paper analyzes the latest research progress in underwater visual SLAM, including the application of multi-sensor fusion, deep learning techniques, and optimization algorithms, which have enhanced the robustness and accuracy of underwater SLAM systems. Additionally, the paper explores the main challenges faced by underwater SLAM technology and proposes potential solutions, such as improving the accuracy of sensor data, enhancing the real-time and robustness of data fusion algorithms, refining feature extraction and matching methods, and elevating the precision and stability of localization and mapping algorithms. Finally, the paper looks forward to future research directions in underwater SLAM, including the application of new types of sensor technology, artificial intelligence techniques, and the development of multi-robot collaborative SLAM for underwater operations. The aim is to provide a comprehensive perspective on the scientific and technological progress in this field. |
