论文部分内容阅读
总结了基于视觉的同时定位和构图技术(VSLAM)的发展历程、分类以及最新研究成果,从图像预处理、前端图构建、后端图优化三个方面详细阐述了基于关键帧的稀疏VSLAM关键技术,展望了VSLAM的未来发展方向,提出了复杂动态环境下VSLAM与惯性导航两种无源定位方法组成组合导航系统的实现框架,为基于视觉的同时定位和构图技术提供了参考。
This paper summarizes the development, classification and latest research results of vision-based VSLAM, and expatiates the key technologies of key-frame-based VSLAM sparsely from three aspects: image preprocessing, front-end graph construction and back-end graph optimization The future development direction of VSLAM is prospected. The implementation framework of integrated navigation system based on VSLAM and inertial navigation in complex dynamic environment is proposed, which provides a reference for simultaneous localization and composition based on vision.