There are a number of tutorials on youtube which explain the steps of creating a 3D model in Blender via the photogrammetry tools VisualSFM and MeshLab. I have distilled those instructions down to their basic elements via the following shortform guides.
An alternate dense point cloud builder: Clustering Views for Multi-view Stereo (CMVS)
VisualSFM:
- file open multiple images: (jpg)
- compute missing matches
- compute 3d reconstruction
- run dense reconstruction (Yasutaka Furukawa's CMVS/PMVS)
Meshlab:
- File -> open project -> bundle.rd.out, list.txt
- ctrl L for panel
- Delete the model mesh
- File -> import mesh -> models -> option-xxxx.ply
- merge the point clouds, if needed
- edit and remove detritus
- Filters -> point set -> surface reconstruction: poisson (12, 7, 3, 1), or ...
- Filters -> select -> select non-manifold edges, delete selection
- Filters -> texture -> parameterization + teturing from registered rasters [try without 'use distance weight']
- file -> export mesh as -> .obj, texture.png
Blender:
- select cycles render
- file -> import .obj
- object -> transform -> origin to geometry
- uv editor
- load texture
- select object, tab to edit, unwrapped polygons should appear on texture, no manual unwrapping necessary
- tab to leave view
- add material
- use nodes
- node editor
- shift a -> input -> texture coordinate
- shift a -> texture -> image texture
- uv <-> vector