I tried to stitch some images from appollo 16 to a panorama. The problem is that I do not know the focal length of the lens. Is there a way to circumvent this?
1 Answer
No, you need to enter either the field of view (FOV), or the lens and crop factor information.
Luckily for you, for the Apollo mission images, the camera and lens information is well documented. The still photos taken by the lunar module team were taken exclusively by Hasselblad cameras modified to accept 70 mm film backs (56mm frame size, 1:1 image ratio), with 60mm Zeiss Metric lenses.
Calculating the angle of view, we get 50° horizontal angle of view:
$$ 2\arctan\frac{\omega}{2f} = 2\arctan\frac{56}{2\times60} = 50º $$
Alternately, you can just enter the focal length of the lens (60 mm) and the crop factor of the 70 mm Hasselblad. The crop factor is with respect to the diagonal distance of 35 mm film camera, which is 43.3 mm. The diagonal distance of 1:1 ratio 70 mm film is just 1.414 * 56 = 79 mm. So the crop factor is 43.3 / 99 = 0.547.
Taking a few of the Apollo 16 images you linked to on Flickr, and plugging in the 60mm lens, 0.437 crop factor into Hugin1, here is the stitched image I created (without any other editing or adjustments):
Note 1: The original version of this answer used a frame width of 70mm for 70mm film. This was wrong. 70mm film used in Hasselblads had a 56mm image height. The "6x6" format (1:1 aspect ratio) is therefore 56mm x 56mm. As edited, the crop factor for 70mm 6x6 should be 0.547. However, when originally creating the Hugin image in this answer, I used 0.437 (as I had wrongly calculated at the time).
-
\$\begingroup\$ Thank you very much for the detailed explanation, that solves my problem! It seems even when having the correct focal length the program will still not suceed if you include too many images whit not many features (left and right of the section you extracted). But I still wonder, is there no way to stitch panoramas where you do not have the information about the FOV or crop factor/focal length? \$\endgroup\$– flawrCommented Mar 19, 2016 at 17:36
-
1\$\begingroup\$ @flawr There is a way – you just have to put in your best guess for the FOV. \$\endgroup\$– scottbb ♦Commented Mar 19, 2016 at 17:52
-
\$\begingroup\$ @flawr Guess, rinse, and repeat until it looks right. \$\endgroup\$ Commented Mar 19, 2016 at 22:29
-
\$\begingroup\$ Well that is a solution too=) I wish there was a way to estimate the Fov from the correspondence points, but I probably would have to implement that myself. Thank you very much for all your inputs! \$\endgroup\$– flawrCommented Mar 19, 2016 at 22:32