|
JAIST Repository >
School of Information Science >
Conference Papers >
Conference Papers >
Please use this identifier to cite or link to this item:
https://hdl.handle.net/10119/20311
|
| Title: | IL-SLAM: Intelligent Line-assisted SLAM Based on Feature Awareness for Dynamic Environments |
| Authors: | Zhang, Haolan
Nguyen Canh, Thanh
Li, Chenghao
Yang, Ruidong
Ji, Yonghoon
Chong, Nak Young |
| Keywords: | Visual SLAM
Feature-aware mechanism
Lineassissted
Dynamic environments |
| Issue Date: | 2025-12-08 |
| Publisher: | Institute of Electrical and Electronics Engineers (IEEE) |
| Magazine name: | 2025 International Conference on Robotic Computing and Communication (RoboticCC) |
| Start page: | 135 |
| End page: | 140 |
| DOI: | 10.1109/RoboticCC68732.2025.00022 |
| Abstract: | Visual SLAM is crucial for autonomous systems but struggles in dynamic environments. While recent dynamic SLAM systems use geometric constraints and deep learning to remove dynamic features, this creates a new challenge: insufficient remaining point features. Existing solutions continuously introduce additional line and plane features, causing unnecessary computational overhead and potential performance degradation from low-quality features. We propose IL-SLAM with a feature-aware mechanism that quantitatively evaluates point feature adequacy (abundance and distribution) to selectively activate line features only when necessary. This mathematically-grounded decision process minimizes computational complexity while reducing additional noise introduction. Our hierarchical optimization strategy uses line features for tracking and local mapping to improve initial pose estimation, but excludes them from global optimization to prevent long-term drift from low-quality additional features. Extensive experiments on TUM datasets demonstrate substantial improvements in both ATE and RPE metrics compared to ORBSLAM3 baseline and superior performance over other dynamic SLAM and multi-feature methods. |
| Rights: | This is the author's version of the work. Copyright (C) 2025 IEEE. 2025 International Conference on Robotic Computing and Communication (RoboticCC), Naples, Italy, pp. 135-140. DOI: https://doi.org/10.1109/RoboticCC68732.2025.00022. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works. |
| URI: | https://hdl.handle.net/10119/20311 |
| Material Type: | author |
| Appears in Collections: | b11-1. 会議発表論文・発表資料 (Conference Papers)
|
Files in This Item:
| File |
Description |
Size | Format |
|---|
| I-CHONG-N-0226.pdf | | 7903Kb | Adobe PDF | View/Open |
|
All items in DSpace are protected by copyright, with all rights reserved.
|
