We aim at developing a new indoor databank to support the intelligent decision of a mobile robot. Such intelligence decisions may include automatous navigation, security surveillance, customer flow analysis, customer guidance, and facility management as examples. The idea is that whenever there is a decision related to the real-world environment to be made, the robot needs a library or databank to refer to make a meaningful decision and to perform the correct follow-up actions.

 

The problem is that up to this moment, there is no such indoor databank in existence compared to the outdoor world, where overhead satellites, CCTV networks on roads, cameras and radars in vehicles, and many others are supplying sufficient data for rebuilding the real world in digital forms for various usage.

 

We adopt a visual approach, like a human to record the most meaningful and useful information with their eye visions. We develop our visual machine learning (VML) algorithms and train the AI with our 360o images that are annotated in 3D GIS-based building models. Now, the VML can be specialised in fire management, customer flow analysis, and autonomous navigation.