Comparison of orb and sift methods under the slam framework

Abstract

This paper presents a comparative study of ORB and SIFT feature matching methods under the SLAM framework. Both algorithms were implemented and tested on identical image sets to evaluate their performance from multiple classical perspectives, including feature distribution, robustness to illumination and rotation, matching accuracy, and computational efficiency. By Analyzing both qualitative and quantitative results, this study aims to provide insights into algorithm selection and optimization for real-time visual SLAM Systems.