Geometry‐guided semantic segmentation for post‐earthquake buildings using optical remote sensing images