Edge SLAM: Edge Points Based Monocular Visual SLAM

Soumyadip Maity, Arindam Saha, Brojeshwar Bhowmick; Proceedings of the IEEE International Conference on Computer Vision (ICCV), 2017, pp. 2408-2417

Abstract


Visual SLAM shows significant progress in recent years due to high attention from vision community but still challenges remain for low textured environments. Feature based visual SLAMs break down due to insufficient features in low textured environment. This paper presents Edge SLAM, a feature based monocular visual SLAM which alleviates this problem. Our proposed method detects edge points from images and tracks those using optical flow, subsequently initialized with robust map quantification. Our method identifies the potential situations where estimating a new camera is becoming unreliable and we adopt a novel method to incorporate the new camera into existing reconstruction using a local optimization technique. We present an evaluation of our proposed system with most popular open datasets. Experimental result indicates that proposed method has comparable accuracy in featured environment and out performed in low textured environment compared to existing monocular SLAM approaches.

Related Material


[pdf] [arXiv]
[bibtex]
@InProceedings{Maity_2017_ICCV,
author = {Maity, Soumyadip and Saha, Arindam and Bhowmick, Brojeshwar},
title = {Edge SLAM: Edge Points Based Monocular Visual SLAM},
booktitle = {Proceedings of the IEEE International Conference on Computer Vision (ICCV) Workshops},
month = {Oct},
year = {2017}
}