Scholarlink

The objective of this project is to create a strong SLAM system aimed at dynamic environments having real-time object detection and tracking incorporated in it. In difficult changing scenes, the system using deep learning will help to improve the accuracy of localization and uniformity of maps. This set out proposal contains among others improvements feature extraction for visual odometry, an object detection framework in addition to adaptive map maintenance. Evaluation shall primarily consider objective criteria such as localization error and presence of the map with various data. Many applications are involved in robotics, augmented reality (AR) an example being Inertial Measurement Unit (IMU) based systems (Kostavelis, 2015), autonomous navigation etc. which need precise localization under highly dynamic real world conditions.

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

Enhancing Visual Simultaneous Localization and Mapping (VSLAM) Robustness in Dynamic Environments

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Tushar Pandita,
  • Vatsal Agarwal,
  • S. J. Tashwin,
  • S. S. Tejas,
  • V. R. Badriprasad

摘要

The objective of this project is to create a strong SLAM system aimed at dynamic environments having real-time object detection and tracking incorporated in it. In difficult changing scenes, the system using deep learning will help to improve the accuracy of localization and uniformity of maps. This set out proposal contains among others improvements feature extraction for visual odometry, an object detection framework in addition to adaptive map maintenance. Evaluation shall primarily consider objective criteria such as localization error and presence of the map with various data. Many applications are involved in robotics, augmented reality (AR) an example being Inertial Measurement Unit (IMU) based systems (Kostavelis, 2015), autonomous navigation etc. which need precise localization under highly dynamic real world conditions.