A robust and efficient mapping and positioning system

We designed a lightweight SLAM system for mapping and positioning in complex environments. The real-time mapping and localization (SLAM) framework we built is mainly divided into two parts: front-end optimization, and back-end closed-loop detection. In the first part of the front-end optimization, a random sampling consensus algorithm is used for ground segmentation and adding ground constraint is to prevent map from bending. In the second part of closed-loop detection, an improved text scanning algorithm is used to optimize the back-end pose. The entire SLAM framework is highly adaptable to the scene, has the characteristics of lightweight and high robustness.