How much time is needed for aligning images that are provided in repository ?

[Khumayun]

How much time is needed for aligning images that are provided in repository ?
I am following step by step workflow to align images, but when it comes to

'Alinging images. Depending on settings this can take from a few seconds to many minutes'

there appears 4 same windows as from capture.plot_undistorted_reflectance(panel_irradiance). I tried waiting as long as I could, but nothing happened. Please, provide with appropriate solution.

P.S. Can't upload screenshot, it says something went really wrong, and we can't process that file,

[maykef]

This will pretty much depend on your system's configuration.
Can you elaborate a bit more on that? What platform are you using, Python version, Opencv installation, etc?
Also, most of these tutorials are informative, is not meant to be a 'support' type page/service. Bear that in mind when open new issues.

[Khumayun]

Processor Intel® Core™ i3-3110M CPU @ 2.40GHz × 4
Graphics Intel® Ivybridge Mobile
OS type 64-bit
OS ubuntu 16.04 LTS
Python 3.5.2 (default, Nov 23 2017, 16:37:01)
[GCC 5.4.0 20160609] on linux
requirements.txt

Ok, I will take into an account. But those multiple windows occurring at the same time, make me confused.

[poynting]

Hi @Khumayun,

Using the code and settings in Alignment.ipynb, it finishes on my core i7 in about 15 seconds.

At one point during development, I had some troubles with a similar problem. The 'fix' on my system was to add the line

multiprocessing.set_start_method('spawn')

to imageutills.py align_capture method. This fixed the issue for me. The symptom I could see in the debugger was that the cv2.findTransforECC call would never complete, and it was clear from CPU usage that it was not doing anything.

If this isn't working, one simple solution could be to just unwrap the enumerate(pool.imap_unordered() call into a series of calls, to remove the multi-processing component entirely. This approach worked for me in all cases, but is a little slower as it doesn't use all CPUs efficiently.