Real-Time SLAM with Point-Line Features Using EDLines in Unknown Aquatic Environments

Autonomous navigation of autonomous surface vessels (ASVs) in unknown aquatic environments is challenging. This paper proposes a real-time point-line feature-based visual simultaneous localization and mapping (SLAM). To solve the problems of inaccurate localization and frequent tracking losses of ASV in poor textural environments, the proposed system introduces line features due to their ability to overcome the interference caused by textureless areas. Futhermore, the edge drawing lines (EDLines) algorithm is applied to extract more efficient line segments in less time. The proposed method is evaluated on both public EuRoc and real outdoor datasets. Experimental results show that the designed system can improve accuracy while ensuring real-time performance.