What OS are you using? Are you au fait with using PHP on the command line? You can definitely cut the overlap back as the images are all well matched for exposure and there's no vignetting. Multiblend doesn't actually need *any* overlap to blend images, but I'd say 5% minimum to be safe, or 10% to be a good balance between blending and number of images needed. On Wednesday, 7 April 2021 at 14:54:00 UTC+1 [email protected] wrote:
> Here is the whole image set so you can pick and choose (may still be > uploading for a while), this is 123 rows x 66 columns ordered left to > right, the focus was not perfect in all shots, expecially the trees on the > right hand side of the bottom few rows, a lens like this has a depth of > feild of about 50m at 1km, > > https://ln.sync.com/dl/4d2d48830/fe4ajnhj-6qy45ytu-catktz9g-w6gusyk8 > > I'm in the process of modifying the mount to be a little more wind > resistant to reduce how much it is able to shift, to hopefully cut back the > overlap % a little, and making my canon camera play nice with the focuser > on the lens, as this batch was manually focused, > > When that is resolved, I'm likely going to try it again, aswell as noting > the exact feild of veiw so I can feed it in to the stitching program > On Wednesday, 7 April 2021 at 8:45:08 pm UTC+10 Monkey wrote: > >> Gigapxtools looks to be an extremely piece of software. As far as I can >> see it just splits or tiles images, with no alignment, overlap, or blending. >> >> ryanf, could you upload a few of your test images somewhere? Say, a >> square of 9, or maybe 25? I'm thinking you might need something more >> bespoke to achieve your goal, which I'd be interested in helping with, as >> I'd love to be able to confirm that Multiblend can blend a terapixel image >> (it should take somewhere on the order of days, and you'll need a few dozen >> terabytes of storage). The fact that the images are taken with such a long >> - and, one assumes, extremely good - lens, and with a motorised mount, >> should make a lot of the usual Hugin workflow redundant and we go straight >> to playing directly with images. >> On Tuesday, 6 April 2021 at 21:51:35 UTC+1 [email protected] wrote: >> >>> Monkey, its 0.44x0.67 degree per image, or about 0.3 square degrees. the >>> total image is going to be roughly 180x90 or 270x60, though I will have >>> some margin on the sides, the goal is to exceed 1 Terapixel >>> >>> GnomeNomad, Thanks I'll check that out, >>> >>> On Wednesday, 7 April 2021 at 3:22:08 am UTC+10 GnomeNomad wrote: >>> >>>> Hmmm, maybe something here could handle? >>>> >>>> https://gigapxtools.com/ >>>> >>>> On April 6, 2021 6:24:42 AM HST, Monkey <[email protected]> wrote: >>>> > Well that's ambitious! If you know how the images overlap, one could >>>> > call >>>> > cpfind manually on each valid image pair just to get the control >>>> > points for >>>> > that pair, then combine all the pairs' control points into one big >>>> > project >>>> > file. I think actually using the Hugin GUI might be out of the >>>> > question - >>>> > unless you replace the images themselves with 1x1 pixel dummies, and >>>> > even >>>> > that might not work - but you might be able to use the command line >>>> > tools >>>> > by themselves to optimise and remap the images. How long it would >>>> take >>>> > is >>>> > another question entirely... >>>> > >>>> > So, 100,000 images with a 1900mm lens... is the end result to be >>>> > 360x180, >>>> > with each image covering a little under a square degree? >>>> > On Tuesday, 6 April 2021 at 14:02:39 UTC+1 [email protected] wrote: >>>> > >>>> > > Would any of you know any good software (free or otherwise), to put >>>> > > together my small scale trial pano of 8118 photos, (Full scale is >>>> > hoped to >>>> > > be 70,000-110,000) shot on a 1900mm lens with a crop factor of 1.62 >>>> > for >>>> > > those interested >>>> > > >>>> > > Hugin seems to crash on trying to open that many files, and seems >>>> to >>>> > lack >>>> > > an easy way to tell it the positions of each of the photos before >>>> > aligning >>>> > > when a smaller subset is used (was taken on a computerised mount) >>>> > > >>>> > > Image composition Editor will begrudgingly handle it, but the >>>> export >>>> > is >>>> > > either limited to less than 65535 in any dimension due to windows >>>> > .NET >>>> > > limitations, or the adobe output that after 3 days was 1% done, and >>>> > lacked >>>> > > means to fix a few stitching errors, (Have looked into patching the >>>> > .NET >>>> > > libraries without much success) >>>> > > >>>> > > I have 256GB of RAM, and many Terabytes of fast storage, so >>>> > computational >>>> > > resources are not an issue, only software that will actually deal >>>> > with it, >>>> > > as this was a 1/10th scale test run before I invest the time in the >>>> > full >>>> > > scale ones, I'm hoping someone e.g. dealing with night sky mosaics >>>> > of sky >>>> > > surveys might have a suggestion, or some means of making hugin work >>>> > for it? >>>> > > >>>> >>>> >>>> >>>> -- >>>> David W. Jones >>>> [email protected] >>>> wandering the landscape of god >>>> http://dancingtreefrog.com >>>> >>>> Sent from my Android device with F/LOSS K-9 Mail. >>>> >>> -- A list of frequently asked questions is available at: http://wiki.panotools.org/Hugin_FAQ --- You received this message because you are subscribed to the Google Groups "hugin and other free panoramic software" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/hugin-ptx/ffb7eec4-5668-41ba-87e0-b245d7fbe947n%40googlegroups.com.
