NV has implemented the process from images to automatic modeling, but the 3D models they produce are architectural systems, not GIS models, can GeoAI achieve similar functions, while fixing the size and geographical location. I have learned a part of GeoAI, but I feel that the main functions still remain in pattern recognition and supervision classification, which is not intelligent enough in my personal feeling. There is neither logical reasoning nor semantic expression, which is far from AI.
