Topological map as an abstract representation of the observed environment has the advantage in path planning and navigation. Here we proposed an online topological mapping method, 360ST-Mapping, by making use of omnidirectional vision. The 360° field of view allows the agent to obtain consistent observation and incrementally extract topological environment information. Moreover, we leverage semantic information to guide topological places recognition further improving performance. The topological map possessing semantic information has the potential to support semantics-related advanced tasks. After combining the topological mapping module with the omnidirectional visual SLAM, we conduct extensive experiments in several large-scale indoor scenes to validate the effectiveness.

System Architecture

Main Video

Supplemental Videos

The topological mapping process in Robotics Institute (an indoor office scenario) and Academic Building (a large-scale multi-space indoor scenario), respectively.

The robustness tests. The left video shows how the system Identifies the same place after the scene has changed greatly (lighting, object layout, object type, etc.) in the Robotics Institute. The video on the right shows the topological mapping result in the Academic Building when the camera's field of view is often blocked during the process.