TY - GEN
T1 - Research on Improving the Robustness and Positioning Accuracy of Visual SLAM Based on Point-Line Feature Matching
AU - Li, Ming
AU - Gu, Shumin
AU - Hou, Xinghua
N1 - Publisher Copyright:
© 2024 IEEE.
PY - 2024
Y1 - 2024
N2 - In order to solve the problem of the failure of SLAM algorithm due to insufficient feature information of a single point in low texture environments, an indoor visual/inertial localization algorithm combining point-line feature matching is proposed. By optimizing the LSD line feature extraction algorithm, the close fusion of vision and IMU data is achieved, and an adaptation factor is introduced in the optimization objective to reduce the processing time. This method reduces the absolute trajectory error and improves the positioning accuracy and system robustness.
AB - In order to solve the problem of the failure of SLAM algorithm due to insufficient feature information of a single point in low texture environments, an indoor visual/inertial localization algorithm combining point-line feature matching is proposed. By optimizing the LSD line feature extraction algorithm, the close fusion of vision and IMU data is achieved, and an adaptation factor is introduced in the optimization objective to reduce the processing time. This method reduces the absolute trajectory error and improves the positioning accuracy and system robustness.
KW - point and line features
KW - simultaneous localization and map building
KW - tight coupling
KW - visual inertia fusion
UR - https://www.scopus.com/pages/publications/105015997946
U2 - 10.1109/IAECST64597.2024.11118090
DO - 10.1109/IAECST64597.2024.11118090
M3 - 会议稿件
AN - SCOPUS:105015997946
T3 - 2024 6th International Academic Exchange Conference on Science and Technology Innovation, IAECST 2024
SP - 1151
EP - 1154
BT - 2024 6th International Academic Exchange Conference on Science and Technology Innovation, IAECST 2024
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 6th International Academic Exchange Conference on Science and Technology Innovation, IAECST 2024
Y2 - 6 December 2024 through 8 December 2024
ER -