This paper develops a multi-dimensional geometric features extraction method for monocular Simultaneous Localization and Mapping(SLAM), in order to meet the self-navigation need for light- weighted robots, e.g. small drones. The monocular SLAM mapping based on feature points method is vulnerable to noisy samples, and thus must be enhanced for better efficiency in complex environments. Our proposed method introduces the line and plane features to the three-dimensional map building process, so as to improve the speed of monocular SLAM applications system's key frames matching and overall stability. In particular, we develop a rapid line matching algorithm, where three-dimensional lines were drawn by two- dimensional lines matching. Numerical results show that our approach can significantly reduce redundant information in the SLAM applications.
A Self-Navigation Method with Monocular Plane Discovery
2017-09-01
417873 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
SYSTEM AND METHOD FOR SELF-SUPERVISED MONOCULAR GROUND-PLANE EXTRACTION
Europäisches Patentamt | 2023
|SYSTEM AND METHOD FOR SELF-SUPERVISED MONOCULAR GROUND-PLANE EXTRACTION
Europäisches Patentamt | 2021
|System and method for self-supervised monocular ground-plane extraction
Europäisches Patentamt | 2023
|Autonomous Navigation Using Monocular ORB SLAM2
Springer Verlag | 2019
|