SfM from Archival images with inconsistent lighting

[danuhl]

Hi all, for a geomorphology project I'm attempting to build 3D models of various mountain faces to compare with modern LIDAR scans. The only success I've had is with images taken very close in time from aerial sources or on the ground, but with consistent lighting and quality.

However most of the archive I'm working with the images are of decent quality, however the light was different in each. I'm stuck with a number of points. If anyone has experience with this: preprocessing in photoshop, strategies for working with difficult images (for example using modern photos to help alignment, or other lidar-generated point clouds. Any advice could help.

In the best case so far the images are aligned with manually point picking, but then the resulting model is basically 2D and very noisy. Could this be due to the way I am processing and saving in photoshop? For example flattening the images? There is no camera data, they are all scanned negatives or prints in tiff format, black and white, some color, however I'm sure I need to convert the color to b&w anyways.

Example images are attached below.

Thanks for any tips!

[3create]

Hi danuhl,

I've worked on projects with archival images: very challenging!
Basically, SfM alone won't get you very far, there are usually too many textural changes between archival and current images for dense stereo matching to work.
The main work is a manual process, not suited for sfm apps.

The first step is to know more about the intrinsic parameters of the archival images. This can be done by matching well spread points of the old and new images (with known coordinates, i.e. from laser scans), a process known as "reverse calibration".

A  well known "classic" in photogrammetry is the reconstruction of the destroyed Buddha statues, maybe that will give you a few hints:
https://www.researchgate.net/publication/227635047_Photogrammetric_Reconstruction_of_the_Great_Buddha_of_Bamiyan_Afghanistan

Guy