Multi-Agent Visual SLAM for 3D Coastline Reconstruction: Overcoming Synchronization and Dynamic Environment Challenges

3D coastline reconstruction presents a significant challenge due to the complex and dynamic environment of coastal areas, characterized by fluctuating sea levels, vegetation, and irregular terrain. Visual Simultaneous Localization and Mapping (V-SLAM) technology offers a promising approach to addressing these challenges by enabling real-time mapping and pose estimation. Conventional V-SLAM methods, which are based on single-agent, are inadequate to meet the complex requirements of such environments. Consequently, multi-agent V-SLAM has emerged as a promising alternative; however, its implementation remains challenged by several unresolved issues. This paper examines the difficulties associated with multi-agent V-SLAM concerning synchronization and points out potential solutions to enhance its efficacy. Additionally, 3D coastline reconstruction is used as a case study to demonstrate the challenges encountered in real-world applications of multi-agent V-SLAM and to propose strategies for overcoming these obstacles.