问题
This question here addresses how to generate a Gaussian kernel using numpy. However I do not understand what the inputs used kernlen and nsig are and how they relate to the mean/standard deviation usually used to describe a Gaussian distribtion.
How would I generate a 2d Gaussian kernel described by, say mean = (8, 10) and sigma = 3? The ideal output would be a 2-dimensional array representing the Gaussian distribution.
回答1:
You could use astropy, especially the Gaussian2D model from the astropy.modeling.models module:
from astropy.modeling.models import Gaussian2D
g2d = Gaussian2D(x_mean=8, y_mean=10, x_stddev=3, y_stddev=3) # specify properties
g2d(*np.mgrid[0:100, 0:100]) # specify the grid for the array
来源：https://stackoverflow.com/questions/42078840/generate-a-gaussian-kernel-given-mean-and-standard-deviation