I want to export the point cloud which is displayed in the viewer after the image alignment. I believe this is also the result of SFM algorithm in the traditional MVS pipeline.
However, I got a very noising point cloud when I export it through “alignment” -> “export” -> “point cloud”. And this is different from the point cloud I saw in the RC viewer.
I do search in the forum and found some relevant poses, but I do not get the answer to my question. Could anyone give me some hints about that?
Also, I am curious about noise points in the sparse point cloud. The reprojected error of these points might be higher than the threshold. Why didn’t RC filter out these points in the SFM pipeline?
Yes, I just want to export the sparse point cloud for research purposes. RC shows that there are 4,406,282 points in the component, but I got 3,539,746 points in the output ply file. The “max displayed point number” is 10,000,000 in the global settings.
That is pretty common during our previous experience with RC. So we try to figure out the reason now. The reconstruction works well though.
Also, our RC version is 1.2.0.16813, if this is useful.