文献类型：期刊文献

英文题名：A Visual-Inertial Slam Method for Robots on Resource-Constrained Platforms Using a Low-Cost Stereo Camera

作者：Liu, Shuang[1]; Yang, Bochun[1]; Tang, Qitao[1]; Li, Songhao[1]; Wei, Hu[1]; Yuan, Hang[1]

机构：[1] School of Mechanical and Power Engineering, East China University of Science and Technology, Shanghai, 200237, China

年份：2025

外文期刊名：SSRN

收录：EI(收录号：20250020492)

语种：英文

外文关键词：Digital storage - Robots - Stereo image processing - Stereo vision

摘要：With the advancement of computer vision, SLAM algorithms utilizing cameras as the primary sensor have become a popular research focus. Many SLAM algorithms based on cameras have demonstrated excellent performance; however, most rely on high-performance hardware. This paper proposes a Visual-Inertial SLAM Method for Robots on Resource-Constrained Platforms Using a Low-Cost Stereo Camera. Firstly, we introduce a localization module that optimizes feature points and performs loop closure detection based on grayscale images, depth maps, and IMU data. Secondly, we propose a map correction module that integrates backend and loop closure detection localization data with depth maps. Finally, we conduct comparative experiments using open-source datasets and real-world experiments on a hardware platform built with low-performance, low-cost equipment. The results demonstrate our approach's accuracy, robustness, and feasibility to low-performance and low-cost devices, highlighting its potential for commercialization and practical applications. ? 2025, The Authors. All rights reserved.