Fix image comparison following numpy 1.21.0 release

[ericpre]

I could not find what changes, but numpy 1.21.0 changes the log norm of matplotlib and it makes one of the plotting test fail. I had to change the test and the reference images so that it works for both version of numpy. The new behaviour is correct: zero values are masked when taking the log norm.

Progress of the PR

• Update test to pass with numpy < 1.21 and numpy 1.21
• add an changelog entry in the upcoming_changes folder (see upcoming_changes/README.rst),
• Check formatting changelog entry in the readthedocs doc build of this PR (link in github checks)
• add tests,
• ready for review.

Minimal example of the bug fix or the new feature

import hyperspy.api as hs
import numpy as np

s = hs.signals.Signal2D(np.arange(10*10).reshape(10, 10))
s -= 50
s.plot(norm='log')

[codecov]

Codecov Report

Merging #2786 (5eccdb2) into RELEASE_next_patch (7fed9b2) will not change coverage.
The diff coverage is n/a.

@@                 Coverage Diff                 @@
##           RELEASE_next_patch    #2786   +/-   ##
===================================================
  Coverage               77.36%   77.36%           
===================================================
  Files                     202      202           
  Lines                   30105    30105           
  Branches                 6579     6579           
===================================================
  Hits                    23290    23290           
  Misses                   5067     5067           
  Partials                 1748     1748           

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 7fed9b2...5eccdb2. Read the comment docs.