Drawing on a Matplotlib Plot | J. Marcus Hughes

It’s easy to draw on a Matplotlib plot and capture the coordinates of the points by using a LassoSelector. Here’s an example script:

import matplotlib.pyplot as plt
from matplotlib.widgets import LassoSelector
import numpy as np

# function saying what to do with the vertices
def on_lasso(vertices):
    print(np.array(vertices))

data = np.random.random((500, 500))

# properties of the line to visualize what you just drew
line_props = dict(color="red", linewidth=2)

# make the visualization
fig, ax = plt.subplots()
lasso = LassoSelector(ax, on_lasso, props=line_props)
ax.imshow(data, vmin=0, vmax=1)
plt.show()

Instead of merely printing the vertices, you could save them to a list for later usage. Or you could get fancy and draw them back on the plot.

I discovered this handy trick when writing SolarAnnotator, a tool to create ground truth data for solar image segmentation.

Overall, the LassoSelector is a powerful way to enable easy annotation in a Matplotlib plot.