while there were few points corresponding to the ground between buildings because the ground was basically in the shadows of the buildings. Therefore, the heights of the interpolated points within the ground area in the DEMs were affected by their neighboring points that were mostly on the buildings. As a result, if there were fewer points on the ground, the heights of the ground between the buildings would be closer to those of the neighboring buildings after interpolation, and there would be no distinct boundary between buildings and the ground in both DEMs. However, more points on the ground in DEM generation would much better describe the height variation of the ground between the buildings in both DEMs. The results can be seen in Fig. 10b and c. Hence, clear boundaries between the collapsed buildings and the ground are observed in the difference between the pre-seismic DEM created from the topographical map and the post-seismic DEM created from the stereo images, while there are no boundaries in some