Initialization is one of the key problems to ensure the localization accuracy of SLAM system, and its performance determines the accuracy of SLAM system to a large extent. However, for the external parameter matrix, ORB SLAM3 still uses the off-line calibration method, once the carrier is affected by external physical impact or mechanical structure mutations, it will inevitably lead to the failure of off-line calibration, and then affect the accuracy of SLAM system. To solve this problem, this paper proposes a full-parameter online initialization algorithm based on monocular ORB SLAM3, which can initialize all parameters including external parameter matrix online when the initial value is unknown, and verifies the effectiveness of the algorithm on the public data set.

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

Full-Parameter Online Initialization Algorithm Based on Monocular ORB SLAM3

  • Jinrong Fan,
  • Longyun Chi,
  • Xujun Guan,
  • Mingrui Hao

摘要

Initialization is one of the key problems to ensure the localization accuracy of SLAM system, and its performance determines the accuracy of SLAM system to a large extent. However, for the external parameter matrix, ORB SLAM3 still uses the off-line calibration method, once the carrier is affected by external physical impact or mechanical structure mutations, it will inevitably lead to the failure of off-line calibration, and then affect the accuracy of SLAM system. To solve this problem, this paper proposes a full-parameter online initialization algorithm based on monocular ORB SLAM3, which can initialize all parameters including external parameter matrix online when the initial value is unknown, and verifies the effectiveness of the algorithm on the public data set.