Results not as expected

[fitodic]

Results not as expected

[AHomoudi]

@fitodic Hello:)

Thanks for the helpful tips. Since am doing programmatically for large set of data, I will leave these steps in case someone will need it :

1. Chaikin's corner-cutting algorithm with refinements degree of X0
2. Shapely simplify function with a tolerance of X1
3. Centerline function with interpolation distance X2

X0, X1, X2 values depend mainly on details of the work and desired results.

[fitodic]

From what I can see, this polygon data set is very detailed. Nevertheless, there are several options you can look into:

1. Modify the border density parameter (by default it is set to 0.5 meters)
2. Simplify the polygon data set (e.g. QGIS)
3. Simplify the centerline data set (e.g. QGIS)

[feihugis]

@fitodic Thanks for your suggestion! The results got much better by simplifying the polygon data using Shapely simplify function.

[fitodic]

Glad I could help 🙂

[AHomoudi]

I'm facing the same issue plus am quite new to Python thus I'm using the function of Centerline inside R. I have tried Shapely simplify function but still have extra lines at the end as shown in the figure. Any ideas for more improvement? Or how to apply the 1 & 3 option?

EDIT: simplifying centerline data set does not remove extra lines at the ends.

[fitodic]

@ahmathlete Hi :)
As to option 1, there is a parameter that increases or decreases the density of the input geometry:

>>> from shapely.geometry import Polygon
>>> from centerline.geometry import Centerline

>>> polygon = Polygon([[0, 0], [0, 4], [4, 4], [4, 0]])
>>> attributes = {"id": 1, "name": "polygon", "valid": True}

>>> centerline = Centerline(polygon, interpolation_distance=1.5, **attributes)

You can also pass the same value to the command line script.

As to option 3, you create the centerline dataset (e.g. using the create_centerline CLI command), open the centerline dataset in QGIS and manually delete the nodes that stick out.

[fitodic]

@ahmathlete Thanks for clarifying 🙂
If you're willing, you can always update the documentation with this info. I know it depends on the dataset, but it never hurts to supply an example.