Research on Loop Closure Detection Based on Deep Learning
LUO Shunxin
ZHANG Sunjie
Abstract：Aiming at the problem of pose drift caused by the positioning and mapping process of mobile robots,the reasons for the pose drift in the whole visual SLAM framework are studied. In the process of moving the mobile robot using the vision sensor,the pose change of the robot is calculated by the pictures of two adjacent frames. As the time increases,the number of picture frames in?creases,and the posture calculated each time is based on the former picture,so the pose error of the robot gradually increases and drifts. The purpose of this paper is to solve the drift problem caused by the robot movement. In this paper,the deep learning method is used to realize the loop detection function to solve the pose drift problem of the robot,and the effectiveness of the method is veri?fied by experiments.
Keywords：visual SLAMloop closure detectiondeep learningpose drift
Publication Date：2019-01-01
Online Publishing Date：2025-08-15(First online date of this platform, not the publication date of the document)
Pages：6( 497-502 )