Point-line feature-based vSLAM systems: A survey

Abstract

The point-line feature-based vSLAM technology significantly enhances the accuracy and robustness of localization and mapping in complex environments by comprehensively utilizing both point and line geometric information. This paper provides a comprehensive survey of methods and applications for point-line feature-based Simultaneous Localization and Mapping (SLAM) systems. Firstly, it focuses on the core components of the visual frontend in SLAM systems, with a detailed analysis of line feature detection methods and their descriptors, covering both traditional algorithms and learning-based approaches, as well as further improvements to these methods. The paper also discusses several common line feature parameterization methods and different line feature matching strategies. In addition, the paper delves into the backend optimization and loop closure detection mechanisms of SLAM systems, which are critical factors in enhancing the system’s accuracy and robustness. By reviewing these methods and applications, this paper aims to provide a comprehensive understanding of integrated point-line SLAM systems, analyzing the strengths and weaknesses of different technologies, and exploring potential directions for future research. This work offers theoretical foundations and practical guidance from a global perspective for the subsequent design and optimization of SLAM systems.