Riassunto analitico
This thesis presents a comprehensive comparison of two image-based SLAM methods, focusing on their performance across different scenarios, including varying lighting conditions, environment complexity and motion dynamics.
The thesis evaluate classical methods like ORB-SLAM3 alongside more recent techniques like Gaussian Splatting SLAM, analyzing their robustness, scalability, accuracy and computational efficiency. Experiments are conducted on standard dataset such as TUM and Kitti visual odometry, to benchmark the methods.
The comparison aims to guide the selection of SLAM methods for specific tasks, contributing to the development of more robust and efficient SLAM systems in robotics and autonomous navigation.
|
