|
Forum: Digital Processing, Software, and Printing
11-07-2018, 12:22 PM
|
| |
I see, I'll look into it if I can bundle the library with the script as well.
|
|
Forum: Digital Processing, Software, and Printing
11-06-2018, 05:47 AM
|
| |
GPU support is only enabled for hugin, I'm pretty sure libvips is CPU only.
Can you try updating libvips to the latest version? (8.7.0). i.e. not pyvips which is bundled with the bin.
|
|
Forum: Digital Processing, Software, and Printing
11-05-2018, 12:32 PM
|
| |
That does look like vips is encountering an error. Do you get any error message or can you reproduce the same error?
Random thoughts: Enough HDD, enough RAM, files aren't removed/renamed/deleted.
|
|
Forum: Digital Processing, Software, and Printing
11-04-2018, 12:38 PM
|
| |
Hmmm curious. Under no circumstances should it take this long. If there isn't an update within a few minutes something went wrong. As hugin already worked for you before and I haven't changed anything in the hugin code this is curious indeed.
This seems to be a hugin error, after "Run called" it should within a few seconds starting with the remapping. The python error is in this case a false positive. Since hugin never generated the aligned images, python can't find them and consequently says that first item in the filelist is out of range as the list of files is empty.
Did you use the same image-stack as before?
---------- Post added 11-04-18 at 12:50 PM ----------
P.s. I tried 8 & 16bit K1-sized images without problems.
|
|
Forum: Digital Processing, Software, and Printing
11-04-2018, 07:08 AM
|
| |
Git repo now has a new file called Superresoltuion.bin this should be standalone executable.
|
|
Forum: Digital Processing, Software, and Printing
11-04-2018, 03:42 AM
|
| |
Ah ok, well stay tuned, next I want to try the python packaging management that should allow to create a package that includes all libraries it depends on without the need to install anything further. So if you get hugin and libvips through fedoras package manager (or build them from git) chances are good you'll get it working as well.
So far I'm really pleased with the results myself!
|
|
Forum: Digital Processing, Software, and Printing
11-03-2018, 08:09 AM
|
| |
I'm sorry to hear this, however in the latest version only pyvips is needed. Give pip3 a try, you can install the packages for your user only and so far it has worked very well for me.
-----------
New Version!
The latest iteration of the superresolution script does no longer need Pillow, Skimage or Numpy. It is now solely based pypis (& libvips) and a lot faster, at least the python part. Hugin will still take its time.
An improvement from using pyvips is that tiling is no longer necessary. Memory requirements have also gone down drastically. 24MP images can now be safely handled <=4GB RAM, also images can be provided in 8bit or 16bit with the same treatment considering the upscaling algorithm.
New version is up on Github now, please read the new readme on how to install libvips & pyvips.
|
|
Forum: Digital Processing, Software, and Printing
10-31-2018, 03:07 AM
|
| |
For reference, I'm using the following versions:
Pillow==5.3.0
numpy==1.15.3
scikit-image==0.14.1
|
|
Forum: Digital Processing, Software, and Printing
10-30-2018, 03:16 PM
|
| |
Hi Terry,
try:
Code:
$ python3
>>> import skimage If skimage is installed correctly, it won't return anything, otherwise you might get an error about a missing module. If the module is not found, try pip to install skimage.
|
|
Forum: Digital Processing, Software, and Printing
10-30-2018, 03:08 AM
|
| |
The format *.md is the markdown format which allows for a bit fancier formatting, but can still be viewed as plain text. If you take a look at GitHub - Rudgas/Superresolution: Another take at the creation of superresolution images using python and hugin the new and updated readme will be displayed below the files :-)
The file can be opened in any regular text-editor. (And is quite readable)
I see you solved it already and I included the debian install command in the readme file. Though I'm no expert on python package management, I'm suprised pip didn't work out. Could it be that you installed it for python2.7? You should be able to check by runnning
pip2 install --user scikit-image
pip3 install --user scikit-image
And see which one returns the already installed message.
I'm currently working on a different angle for 16bit images to allow bilinear extrapolation so stay tuned. Would be glad for the example images though!
I'm sorry I can't help you with fedora specifically (I'm runnning Arch), but have you tried:
pip3 install --user scikit-image
?
-----
In other news, soon there will be a changelog and versioning so it is a bit easier to see which improvements have been made. Thanks for the feedback so far, keep it coming :-)
|
|
Forum: Digital Processing, Software, and Printing
10-29-2018, 01:05 AM
|
| |
Script is now updated with 16bit capabilities.
To process K1-sized images in 16bit, make sure you have around 6GB of available (free -h) memory.
A minor drawback with 16bit images, is that they have to be split into tiles before resizing wich only allows nearest-neighbor extrapolation rather than bilinear which is used for 8-bit. This is something I'm currently looking into, also I'm not sure yet about the theory of extrapolation methods for the creation of superresolution pictures.
|
|
Forum: Digital Processing, Software, and Printing
10-28-2018, 06:58 AM
|
| |
Actually never mind about the images, found a decent stack.
|
|
Forum: Digital Processing, Software, and Printing
10-28-2018, 04:18 AM
|
| |
Thanks for trying the script!
I added a shebang line that should be compatible across all systems (#!/usr/bin/env python3) which will call the most recent python 3 version.
Should have probably mentioned this, this is a warning that the program is loading a compressed image (tif) into memory, however in memory images are huge because each pixel has to be stored individually for each color channel. This is a warning that the image that is loaded exceeds the expected size and that it might clog your RAM.
Wasn't aware the K1 could do this, will try to implement it. As for now it is 8-bit only.
Yes! I had a rounding error for the tiles, I preallocated an array with the size of the tiles, however, if the enlarged image has an uneven number of pixels along one axis, tiles will not be the same size (from shape 4889 -> 4888). This is now fixed! (no pixels are lost, code is just not as neat as before)
---------- Post added 10-28-18 at 04:22 AM ----------
@jbinpg could you send me the stack of 16bit tif? Can't find a good source for them online and would like to implement the 16bit processing.
|
|
Forum: Digital Processing, Software, and Printing
10-27-2018, 02:53 PM
|
| |
^ And this is why I love pentaxforums :-)
|
|
Forum: Digital Processing, Software, and Printing
10-27-2018, 01:08 AM
|
| |
Hi all,
I browsed through the net for a superresolution script to play around with, however most scripts I found weren't really memory friendly, easily running out of memory with 24MP images and using large stacks.
So I made yet another my own version, using python and hugin. Can process jpg and tif, 8bit and 16bit.
System requirements:
Linux would probably a whole lot easier than using windows.
Hugin (open source panorama software)
Python3
- Python packages: pyvips
For a ~15 Image Stack of 8-bit 24MP Images:
RAM 4GB will be enough for 24MP images
~6-7GB HDD the tif files are huge
Takes ~12min. on my old i3-3225
There is actually no limit on stack-size concerning RAM when using the average method. Using median, at some point memory could run out (not soon).
How it works:
Step 1: Images are resized to 200% using Bilinear extrapolation (tried several, looked best to me), saved as tif (uncompressed to speed stuff up).
Step 2: Images are aligned using hugin, and cropped to an area all images cover. (this is done on the gpu)
Step 3: Average & Median values are calculated for each pixel (each channel RGB separately)
Step 4: The two output files are created and saved in the folder superresolution
How to use it:
Please see GitHub - Rudgas/Superresolution: Another take at the creation of superresolution images using python and hugin for installation instructions. I'll post about updates here in the thread.
(old) Results (Handheld High-Burst with a K3 and a A-50mm/f1.7) below:
|
|
Forum: Pentax K-3 & K-3 II
09-29-2018, 11:13 AM
|
| |
Well, that worked as well on the K30. But my nose is in the way when trying to press the green button when using the viewfinder :--) Seems I have to get the DA 50mm after all :-D
|
|
Forum: Pentax K-3 & K-3 II
09-29-2018, 10:09 AM
|
| |
Just looking at the Lens when taking a photo, I could see the aperture closing. Also, correctly exposed pictures when using the EV compensation. I'm sure I have some pictures with ev +2 or something and correct exposure if absolute proof is needed ;-)
Dammit! I knew it a was a magical bond between me and my K30! Maybe the K3 is just shy?? But kidding aside, the Av-Stopdown works only with a non-conducting base. I discovered this by accident using a Porst-Lens and not being able to get the green button working. With a regular M-Lens, stop down does also not work in Av Mode. So this is why I cover the contacts with a bit of tape to trick the camera.
|
|
Forum: Pentax K-3 & K-3 II
09-29-2018, 09:24 AM
|
| |
Ah yes underexposed. I don't have the camera anymore, so it's a bit confusing.
No, this was no error, the difference is just that the K30 would stop down the lens in Av mode as opposed to the K3 which does not stop down the lens (or at least I can't get it to stop down in Av mode). This is quite a bit disappointing :-/ M-Lenses were just so easy to use with the K30...
|
|
Forum: Pentax K-3 & K-3 II
09-29-2018, 08:24 AM
|
| |
Firmware 1.05, maybe 1.06 (latest).
|
|
Forum: Pentax K-3 & K-3 II
09-29-2018, 07:56 AM
|
| |
With the K30, if there was no electrical connection to the lensmount, the camera would try to stop down the lens as much as set in by the aperture ring. So normally the aperture is held open, but on pressing the shutter, the camera would stop down the lens to whatever the aperture ring was set to.
Now the interesting part is, the K30 would adjust exposure automatically for widest aperture (e.g. f1.7) by itself. Now if I set the aperture ring to f4, the K30 would still measure at f1.7 and stop down the lens (as far as it would go) to f4 prior to taking the shot. This results in a picture that is overexposed by exactly 2.5 steps (slight error in previous post).
Now these 2.5 steps can be set by using the cameras exposure compensation, thus eliminating the need for the green button.
This in turn means that you have to adjust EV comp. with changing the aperture, however this is much less trouble than the classic stop-down metering.
For simplification ev-steps for M 50mm f1.7:
1.7 -> +0
2.0 -> +0.5
2.8 -> +1.5
4 -> 2.5
|
|
Forum: Pentax K-3 & K-3 II
09-29-2018, 05:05 AM
|
| |
Yes, also using Av Mode. In M, stop-down works with the green button, but it's a huge loss of comfort having to press the green button before every shot.
|
|
Forum: Pentax K-3 & K-3 II
09-29-2018, 04:38 AM
|
| |
Hi all,
just got me a new K3, and the first lens I had to try was the 50mm f1.7. However I noticed some odd behavior.
Previously I owned a K30, and for manual lenses, I usually covered the Lens where the contacts would be with some tape. The reason behind this is I could set the aperture to, let's say 4 and adjust exposure compensation by +3.5 (from 1.7-4) and bob's your uncle, no stop down metering (green button) necessary.
However, the K3 does not seem to stop-down the aperture and instead always takes the shot fully open, why is that and is it possibly to change this?
Cheers!
|
|
Forum: Pentax DSLR Discussion
09-28-2018, 02:07 AM
|
| |
I got an offer for a K3 with ~22k shuttercount for 370€ and took it. Seems to me a very good deal considering all alternatives.
K70 was tempting, but too pricey (600€) considering the main benefits were in the IQ department. Which is nice, but with 24" Screen wallpapers and 14" prints so far, I won't get much benefit from it. The only other tempting feature was wifi, but since I haven't used this feature once on my Lumix GX80 during my holiday, where it would actually make sense without a pc around, I can skip it.
K5II was just no competition price-wise.
|
|
Forum: Pentax DSLR Discussion
09-27-2018, 10:55 AM
|
| |
Checking K5II prices, the difference is very small to the K3 ~50-100€ and I think that's worth it for me.
|
