信息与控制2024，Vol.53Issue(4)：499-507,9.DOI:10.13976/j.cnki.xk.2024.3198
融合深度学习与稠密光流的动态视觉SLAM
Dynamic Visual SLAM Integrating Deep Learning and Dense Optical Flow
摘要
Abstract
Traditional visual simultaneous localization and mapping(SLAM)algorithms function well when the environmental objects are stationary or moving at low speeds,but their precision and ro-bustness are low when dynamic disturbances such as personnel walking and vehicle moving are present.To address this problem,a dynamic SLAM system is proposed based on the Oriented FAST and Rotated BRIEF-SLAM3(ORB-SLAM3)framework,which integrates the You Only Look At CoefficienTs(YOLACT++)deep learning network with the ORB-SLAM3 framework for detecting dynamic targets.A dense optical flow field is extracted and incorporated with visual ge-ometry to discover the motion attributes.A motion-level transfer strategy that integrates an instance segmentation network and dense optical flow field to achieve joint optimization of SLAM system ef-ficiency and accuracy is proposed.The test results on the public dataset TUM present that the pro-posed system has an outstanding performance in dynamic scenarios.Compared with ORB-SLAM3,the root mean square error,mean error,median error,and standard deviation in low dynamic sce-narios are enhanced by approximately 60％and over 90％in high dynamic scenarios.The actual experiments in a corridor scene reveal that the proposed system can effectively eliminate feature points on dynamic targets while extracting features,thus guaranteeing system accuracy.关键词
视觉SLAM/动态目标干扰/深度学习/稠密光流/视觉几何Key words
visual SLAM/dynamic target interference/deep learning/dense optical flow/visual geography分类
计算机与自动化引用本文复制引用
胡青松,任林洋,随学帅,李世银,孙彦景..融合深度学习与稠密光流的动态视觉SLAM[J].信息与控制,2024,53(4):499-507,9.基金项目
国家自然科学基金项目(52474185,51874299) （52474185,51874299）
"双一流"建设提升自主创新能力项目(2022ZZCX01K01) （2022ZZCX01K01）
中国矿业大学"工业物联网与应急协同"创新团队资助计划(2020ZY002) （2020ZY002）
