Autonomous navigation method of four-phase wheel foot wing submersible multi-mode intelligent robot

The invention relates to an autonomous navigation method of a four-phase wheel foot wing submersible multi-mode intelligent robot, which adopts a semantic information extraction technology based on deep learning to carry out target detection and semantic information extraction on a dynamic target in the surrounding environment of the robot, and adopts a VI NS dynamic autonomous navigation technology based on semantic recognition to carry out target detection and semantic information extraction on a dynamic target in the surrounding environment of the robot. Performing semantic information labeling and visual odometer construction to obtain a three-dimensional map of the land and air environment around the robot; removing redundancy of the obtained three-dimensional map by adopting a point cloud compression method; an NDT algorithm is adopted for repositioning of the robot in the three-dimensional map with redundant information removed, and global path planning and local path planning are conducted in combinat