feature request: maxdim in CPfind
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Hugin |
Confirmed
|
Wishlist
|
Unassigned |
Bug Description
from this thread:
https:/
Basically the --fullscale flag is not very convenient. Better would be something similar to autopano sift c, which allows the user to specify the image dimension to use for analysis.
Harry had these comments:
I suggest to file it as a feature request.
I assume your idea with the "--maxdim <n>" like in autopanosift-c can be picked up very easily.
for autopanosift-c: "--maxdim <n> Make largest image dimension <= n (default: 1600)."
However, this might fall under the license on APSC.
We also have calibrate_lens which has the option: "-d <int> Maximum dimension for re-sized image prior to processing. Default 1600".
calibrate_lens is GPL.
We also have celeste_standalone with the option: " -s <int> Maximum dimension for re-sized image prior to processing. A higher value will increase the resolution of the mask but is significantly slower. Default: 800"
celeste_standalone is GPL
Finally we have align_image_stack which uses the "-s scale Scale down image by 2^scale (default: 1 [2x downsampling])". This option might not be as good whith varying input resolutions but is also GPL.
Enough options to copy&paste into cpfind.
Changed in hugin: | |
importance: | Undecided → Wishlist |
status: | New → Confirmed |
at least now we no longer have the calibrate_lens variant, but there are definitely too many confusing options. Ideally there should be one option that behaves the same across all tools.
Tom Sharpless had the idea of using the images from the pyramid already used for finding CPs rather than an arbitrary (and costly) rescale.
--pyramid-level N with N indicating the level of the pyramid to be used ?