3D Mapping and Indoor Navigation for an Indoor Environment of the University Campus

Indoor navigation systems help provide route guidance in indoor environments, especially for large and complex spaces. The development of indoor navigation systems includes several components, namely generating the indoor map to be used as the basis of the navigation system, mapping points of interest (POI), indoor positioning to locate the user’s current position and route guidance for providing instructions to guide the user to his/her desired destination. When the indoor space gets larger or more complicated, the conventional manual processes of map creation and POI mapping become tedious and prone to errors. A semi-automated indoor navigation system framework is proposed to solve the mentioned issues. The proposed framework replaces tasks requiring manual operation with automated workflows requiring minimal human intervention. Indoor mapping using a stereo camera is used to reconstruct the 3D model of the indoor environment. The generated 3D model eliminates the need to create the 3D map from the 2D floorplan and is useful for environments with no floorplans readily available. POI mapping is performed using object detection and optical character recognition (OCR) methods. Eventually, a mobile application is developed which utilizes the 3D map, the POI, the OCR results with YOLO detection algorithm, the NavMesh walkable surfaces and A* Pathfinding algorithm to provide path generation and navigation functionalities. Overall, the proposed framework speeds up the process of indoor mapping and navigation significantly as compared to conventional methods.