Agisoft Metashape

Agisoft Metashape => General => Topic started by: notInsane on March 17, 2020, 02:48:05 AM

Title: historic aerial frames, same pixel dimensions, fiducials slightly different
Post by: notInsane on March 17, 2020, 02:48:05 AM

Hello,
I was able to locate some fairly high-quality 1960s-era aerial photograph negatives and had them scanned at 12 micron on a well-calibrated photogrammetric scanner. Metashape Pro 1.6.2 build 10247 detects the fiducial marks very well. However, when testing individual cameras and 'calibrate fiducials' I get some shifts between the different frames.

There is no available calibration data that I have been able to find.

I quick check shows variation between pairs of fiducials on a frame generally less than 0.2 mm, with shifts between frames up to 0.5 mm; the larger shifts appear to be between frames oriented up and down.

So the photos were acquired with the same camera but as scanned the calibration groups perhaps should be different. Is there a problem with having basically one calibration group per frame? In this project I'm using 14 frames but am interested in an area in the middle of the coverage.

Thanks,
KJC

Title: Re: historic aerial frames, same pixel dimensions, fiducials slightly different
Post by: radar_boy on March 27, 2020, 02:15:26 PM

First, which country, agency/company and camera? I have some archive camera calibration reports and you never know, they might match...!

Second, the variations you report (0.2 and 0.5 mm)... are these with reference to the [assumed] principal point?

Andy