未知环境下基于视觉与激光融合的SLAM系统研究与实现

Research and Implementation of SLAM System Based on Vision and Laser Fusion in Unknown Environment

在未知环境中实现机器人的自主导航与环境建图是一个极具挑战的研究课题,SLAM是应对这一挑战的关键技术之一。基于研究现状分析、系统方案设计、技术实现与实验验证等内容,结合视觉与激光各自的优势,设计一种未知环境下基于视觉与激光融合的SLAM系统。通过针对关键帧和滑动窗口的平衡选取策略以及分类优化策略对初始位姿和特征点进行优化,提高了系统的鲁棒性,可以输出高精度地图;通过实现以智能小车为载体的演示系统验证了所提出算法的可行性和实用性。

Realizing autonomous navigation and environmental mapping of robots in unknown environments is a highly challenging research topic,and SLAM is one of the key technologies to address this challenge.Based on the analysis of research status,system scheme design,technical implementation,and experimental verification,combined with the advantages of vision and laser,a SLAM system based on vision and laser fusion in an unknown environment is designed.By using a balanced selection strategy for keyframes and sliding windows,as well as a classification optimization strategy,the initial pose and feature points are optimized,which improves the robustness of the system and can output high-precision maps;the feasibility and practicality of the proposed algorithm are verified through the implementation of a demonstration system using intelligent cars as carriers.