For my research in infrastructure monitoring, I need a 3d model of the infrastructure (just a rough model is fine). I decided to use photogrammetry, so I took some photos of the infrastructure along a baseline. I tried to generate the 3d model using the Agisoft Metashape software because it seems to be the most user-friendly app, and I don't really have a lot of experience in photogrammetric mapping.
I used a general processing scheme (align photos -> build model and point cloud). Somehow, the further part of the observed area, which is the infrastructure part that I want to have, does not get generated. However, the tie points are seemingly quite dense over the infrastructure, and the infrastructure in the photograph did get captured very clearly, so I thought there wouldn't be any problem. here are some pictures for illustration.
One of the photograph to build the model
The resulted tie points after photos alignment
The generated model (same perspective as the tie points picture)
I wonder if it's because Agisoft is more specialized in aerial photogrammetry using drones so it disregards the further parts. Is there any way that I could generate the 3d model (preferably the point clouds) of the bridge?
I took around 70+ photographs facing the bridge along a horizontal baseline perpendicular to the bridge. The bridge is captured very clearly in the photographs and the bridge does not have much complex shapes and structure. So I expected the model of the bridge to be generated well. However, only the area close to the camera was able to be generated as the 3d model.
According to the screenshot it seems that the reconstruction volume (region box) does not include all the area of interest. You need to re-adjust it and re-start the surface generation process (including re-building of the depth maps).
