1f. Local data weighting

LoopStructural primarily uses discrete interpolation methods (e.g. finite differences on a regular grid, or linear/quadratic on tetrahedral meshes). The interpolation is determined by combining a regularisation term and the data weights. The default behaviour is for every data point to be weighted equally, however it is also possible to vary these weights per datapoint.

from LoopStructural import GeologicalModel
from LoopStructural.datasets import load_claudius
from LoopStructural.visualisation import LavaVuModelViewer

Use Cladius case study

data, bb = load_claudius()
data.head()

	Unnamed: 0	X	Y	Z	val	feature_name	nx	ny	nz
0	0	550551.810547	7.821922e+06	-9734.299805	250.0	strati	NaN	NaN	NaN
1	1	550551.810547	7.821903e+06	-9731.809570	250.0	strati	NaN	NaN	NaN
2	2	550551.810547	7.821884e+06	-9730.330078	250.0	strati	NaN	NaN	NaN
3	3	550551.810547	7.821865e+06	-9729.330078	250.0	strati	NaN	NaN	NaN
4	4	550551.810547	7.821847e+06	-9729.679688	250.0	strati	NaN	NaN	NaN

Build model with constant weighting

Build model with weight 1.0 for the control points (cpw) and gradient normal constraints (npw)

model = GeologicalModel(bb[0, :], bb[1, :])
model.data = data
model.create_and_add_foliation("strati", interpolatortype="FDI", cpw=1.0, npw=1.0)
view = LavaVuModelViewer(model)
view.add_isosurface(model["strati"], slices=data["val"].dropna().unique())
view.rotation = [84.24493408203125, -77.89686584472656, -171.99795532226562]
view.display()

Change weights to

model = GeologicalModel(bb[0, :], bb[1, :])
model.data = data
model.create_and_add_foliation("strati", interpolatortype="FDI", cpw=10.0, npw=1.0)
view = LavaVuModelViewer(model)
view.add_isosurface(model["strati"], slices=data["val"].dropna().unique())
view.rotation = [84.24493408203125, -77.89686584472656, -171.99795532226562]
view.display()

Locally vary weights # ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Add a weight column to the dataframe and decrease the weighting of the points in the North of the model.

data, bb = load_claudius()
data["w"] = 1.0
data.loc[data["Y"] > (bb[1, 1] - bb[0, 1]) * 0.2 + bb[0, 1], "w"] = 0.01
data.sample(10)

# cpw/npw are multipliers for the weight column
model.create_and_add_foliation("strati", cpw=1.0, npw=1)
view = LavaVuModelViewer(model)
view.add_isosurface(model["strati"], slices=data["val"].dropna().unique())
view.rotation = [84.24493408203125, -77.89686584472656, -171.99795532226562]
view.display()