Closed kacperbudny closed 3 years ago
Can you provide the image, or the numpy.array of the image, or at least you can provide the numpy.ndim(arr)
output
Hi @kacperbudny,
Beautiful image. It seems that the problem comes because it's a color image with 3 channels RGB. Currently astroalign
only works with grayscale images (1 channel). I suggest picking one channel to find the transformation and then apply it to all the color channels.
This situation is common enough that we should probably include it in the next release. I will create an issue for the feature addition. In the meantime, let me create a gist with a recipe to deal with color images.
something like this (not fully tested)
https://gist.github.com/martinberoiz/1ab5f6ec4a81680517ad2d9bd2ab9f32
I hope it helps.
Hello,
I just released version 2.3.1 that adds support for color images. Please try again with new version and if you find any problem, you can open a new issue.
In the meantime I'll go ahead and close this issue.
Hi. This library seems like just what I needed, but I can't get it to work on the photos I have.
Here's the issue: I load my images using cv2 (imread), so they are numpy arrays, as the documentation requests. But whenever I try to register these photos, I get this:
Traceback (most recent call last): File "...\astroalign.py", line 243, in find_transform source_controlp = _find_sources( File "...\astroalign.py", line 465, in _find_sources bkg = sep.Background(image) File "sep.pyx", line 396, in sep.Background.cinit File "sep.pyx", line 304, in sep._parse_arrays File "sep.pyx", line 232, in sep._check_array_get_dims ValueError: array must be 2-d
During handling of the above exception, another exception occurred:
Traceback (most recent call last): File "main.py", line 13, in
transf, (source_list, target_list) = aa.find_transform(source, target)
File "...\astroalign.py", line
247, in find_transform
raise TypeError("Input type for source not supported.")
TypeError: Input type for source not supported.
Am I doing something wrong?