2017/05/26 by Ruben Gomez-Ojeda, David Zuñiga-Noël, Gomez-Ojeda, Ruben +7 · 5 citations
Computer Science · Engineering · #Advanced Image and Video Retrieval Techniques #Advanced Vision and Imaging #Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences #Robotics and Sensor-Based Localization
paper · pdf · doi:10.48550/arxiv.1705.09479
openalex publication_date 2017/05/26 · openalex created_date 2022/10/02 · openalex updated_date 2026/07/28
Traditional approaches to stereo visual SLAM rely on point features to\nestimate the camera trajectory and build a map of the environment. In\nlow-textured environments, though, it is often difficult to find a sufficient\nnumber of reliable point features and, as a consequence, the performance of\nsuch algorithms degrades. This paper proposes PL-SLAM, a stereo visual SLAM\nsystem that combines both points and line segments to work robustly in a wider\nvariety of scenarios, particularly in those where point features are scarce or\nnot well-distributed in the image. PL-SLAM leverages both points and segments\nat all the instances of the process: visual odometry, keyframe selection,\nbundle adjustment, etc. We contribute also with a loop closure procedure\nthrough a novel bag-of-words approach that exploits the combined descriptive\npower of the two kinds of features. Additionally, the resulting map is richer\nand more diverse in 3D elements, which can be exploited to infer valuable,\nhigh-level scene structures like planes, empty spaces, ground plane, etc. (not\naddressed in this work). Our proposal has been tested with several popular\ndatasets (such as KITTI and EuRoC), and is compared to state of the art methods\nlike ORB-SLAM, revealing a more robust performance in most of the experiments,\nwhile still running in real-time. An open source version of the PL-SLAM C++\ncode will be released for the benefit of the community.\n