The goal of this project is to provide a full semantic mapping solution. It is based on a fusion of deep learning image segmnetation algorithm (Mask-RCNN) and SLAM algorithm (RTAB-Map). Additionally few algorithms are provided for semantic pointcloud processing to prove usefulness of the concept and the solution.
The proposed solution is based on exisiting repositories with rtabmap and Mask-RCNN implementations. For clarity, in this repository, only modified files of these repos are included.