Keyframe-based Dense Mapping with the Graph of View-Dependent Local Maps
This work addresses incremental improvements in robotic mapping and localization for applications like autonomous navigation.
The authors tackled the problem of dense 3D mapping by proposing a keyframe-based system that updates local Normal Distribution Transform (NDT) maps using RGB-D sensor data, achieving improved precision for objects closer to the camera and enabling global map correction after loop closure detection.
In this article, we propose a new keyframe-based mapping system. The proposed method updates local Normal Distribution Transform maps (NDT) using data from an RGB-D sensor. The cells of the NDT are stored in 2D view-dependent structures to better utilize the properties and uncertainty model of RGB-D cameras. This method naturally represents an object closer to the camera origin with higher precision. The local maps are stored in the pose graph which allows correcting global map after loop closure detection. We also propose a procedure that allows merging and filtering local maps to obtain a global map of the environment. Finally, we compare our method with Octomap and NDT-OM and provide example applications of the proposed mapping method.