To use Umbra, you simply need to download the latest release. The executable may be flagged by your antivirus as a false positive. If this happens, consider adding it to your antivirus's exclusion list.
The source code is available if you want to run the scripts without using the GUI.
- Python 3.7+
- Packages in
requirements.txt
The script scripts/registration.py simultaneously performs a moon-based and a sun-based registration of the input images.
Parameters (defined at the bottom of the script) :
input_dir: Input folder containing the FITS files.ref_filename: All images will be both moon-aligned and sun-aligned to this reference image. Ideally, the reference image should be a long exposure with the inner solar corona clipped. It should have the same camera settings as the anchor images.anchor_filenames: The anchor images are the only images that will be explicitly sun-aligned to the reference using the sun registration algorithm. The other images will use timestamp-based interpolation to compute the relative translation of the sun (with respect to the moon), as well as the rotation. One or two anchor images are generally sufficient. They should have the same camera settings as the reference image, and be spaced as far apart in time as possible from it.moon_registered_dir: Output folder for the moon-registered images.sun_registered_dir: Output folder for the sun-registered images.image_scale: Resolution in arcseconds/pixel.clipped_factor: In order to easily detect the moon's border, the bright pixels surrounding the moon are clipped first, if they are not already. This parameter determines the number of clipped pixels. Increase to make the moon's border more defined. Decrease to prevent noise amplification (which may interfere with the edge detection algorithm). The number of clipped pixels is computed as the area of an annulus around the moon, where the outer radius is given by the moon radius, multiplied by the clipped factor.edge_factor: The moon detection algorithm works by fitting a circle to the edge of the moon. This parameter determines the number of detected edge pixels, displayed in red and green. Increase if a large portion of the moon's border is not detected. Decrease if other parts of the image are incorrectly detected. The number of edge pixels is given by the circonference of the moon, multiplied by the edge factor. Some edge pixels are then discarded, due to non-maximum suppression.sigma_highpass_tangential: The sun registration algorithm works on filtered images that enhance the coronal details. This parameter defines the standard deviation of the tangential high-pass filter, given in degrees. A lower value emphasizes finer structures, while a higher value is more robust to noise.max_iter: Maximum number of iterations for the optimization loop. The loop will terminate early if the parameters of the alignment transform converge.