Scholarlink

<p>Visual Simultaneous Localization and Mapping(SLAM) technology is widely used in the autonomous navigation of mobile robots. However, in the face of increasingly complex application scenarios, there are still many problems that need to be solved, among which poor localization in weak texture environments is one of the most important problems. Therefore, in this study, to address the problem of large positioning errors caused by insufficient point features in weak texture environments, based on Oriented FAST and Rotated BRIEF Simultaneous Localization and Mapping 2(ORB-SLAM2), we optimize the estimation of the camera’s position information by extracting the line features in the environment and integrating the point and line features and realize a visual odometry that can simultaneously extract the matching point features and line features, which enhances the positioning accuracy and stability of the SLAM system in weak texture environments. Experiments are conducted on the The Technical University of Munich (TUM) dataset and real scenes, and the experimental results show that the constructed algorithm can effectively utilize the point and line feature information in the environment, and the localization accuracy is improved by about 30% in the indoor structured texture-free scene, about 10% in the indoor structured textured scene, and about 15% in the other indoor scenes. The experimental results demonstrate the effectiveness of the visual SLAM algorithm based on point and line features proposed in this paper.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A visual SLAM algorithm based on point and line features in weak texture environments

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年1月23日

查看全文
  • Rui Li,
  • Dong Yang,
  • Shiqiang Yang,
  • Huandong Li,
  • Xingang Yang,
  • Xiaoli Hou,
  • Fanlin Kong

摘要

Visual Simultaneous Localization and Mapping(SLAM) technology is widely used in the autonomous navigation of mobile robots. However, in the face of increasingly complex application scenarios, there are still many problems that need to be solved, among which poor localization in weak texture environments is one of the most important problems. Therefore, in this study, to address the problem of large positioning errors caused by insufficient point features in weak texture environments, based on Oriented FAST and Rotated BRIEF Simultaneous Localization and Mapping 2(ORB-SLAM2), we optimize the estimation of the camera’s position information by extracting the line features in the environment and integrating the point and line features and realize a visual odometry that can simultaneously extract the matching point features and line features, which enhances the positioning accuracy and stability of the SLAM system in weak texture environments. Experiments are conducted on the The Technical University of Munich (TUM) dataset and real scenes, and the experimental results show that the constructed algorithm can effectively utilize the point and line feature information in the environment, and the localization accuracy is improved by about 30% in the indoor structured texture-free scene, about 10% in the indoor structured textured scene, and about 15% in the other indoor scenes. The experimental results demonstrate the effectiveness of the visual SLAM algorithm based on point and line features proposed in this paper.