How to fill the holes in DSM.tif? about s2p

[dyoussef]

The parameter dsm_radius allows you to interpolate the small holes.
Many small holes can appear when the sensor resolution and the dsm_resolution are similar.

Regards

[thedodo]

Of course no data values can always happen, it does not mean that the output is wrong (for instance through self-occlusion).
Without having the image pair that was used it is hard to say if this is wrong or not.
Does the 3D model make sense if you view it e.g. in QT Reader?

As to how to get rid of them: You can use any available computer vision tool, e.g. interpolate non-existing data using neighborhood information or use convolutions (for instance bilateral/guided filter) either on the disparity map or on the height map directly.

best regards

[lionlai1989]

Thank you. I change the dsm_radius from 0 to 1, and it works.