search

champsproject / ldds

Python package for computing and visualizing Lagrangian Descriptors in Dynamical Systems
BSD 3-Clause "New" or "Revised" License
18 stars 2 forks source link

Total LD output using `max_gradient` implementation #62

Closed broncio123 closed 3 years ago

broncio123 commented 3 years ago

Hi @VikJGG and @vkrajnak

I ran two simple tests.

TEST 1 Before and After max_gradient implementation

image image

TEST 2 After using high_constrast_gradient Similar output to Laplacian-based filter solution with single threshold, but now the user would have to set up two thresholds?

image

vkrajnak commented 3 years ago

What precisely is the bug? Is it the second figure in Test 1 that seems to have fewer contour levels?

vkrajnak commented 3 years ago

If that is the problem, it is a consequence of using max instead of a sum of LD gradients. Does it happen anywhere besides the 1D saddle? Nothing unusual shows up in 2D saddle or Hénon-Heiles.

Can't see how else to make the points near manifolds stand out without consequences. Any ideas?

vkrajnak commented 3 years ago

@broncio123 Is that what you meant by the issue or something else?

vkrajnak commented 3 years ago

No response, I assume this is clear then