import os
from random import sample
from openalea.plantgl.all import (
NurbsCurve2D,
QuantisedFunction,
Scene,
connect_all_connex_components,
densities_from_r_neighborhood,
k_closest_points_from_ann,
r_neighborhoods,
)
from .livnymethod import livny_method_mtg
from .mtgmanip import initialize_mtg
from .sca import adaptivespacecolonization_method, spacecolonization_method
from .xumethod import graphcolonization_method, xu_method
[docs]
def load_points(filename):
"""
Load a point cloud from a file.
Parameters
~~~~~~~~~~
filename: str
Path to the file of the point cloud.
Returns
~~~~~~~
PointList of the point cloud.
"""
# Check if file exists
assert os.path.exists(filename)
s = Scene(filename)
# Check if pointList is on the loaded file
assert len(s) == 1
geom = s[0].geometry
# Unwrap transformations if needed
if hasattr(geom, "geometry"):
# e.g. Translated, Scaled, etc.
points = geom.geometry.pointList
# points.translate(s[0].geometry.translation)
else:
points = geom.pointList
return points
[docs]
def subsample(points, ptsnb=None):
"""
Subsample a point cloud using random.sample algorithm.
Parameters
~~~~~~~~~~
points: PointList
PointList of the point cloud, usualy obtained with :func:`load_points`.
ptsnb: int | None
Number of points to keep.
Default to None, which means that the number of points is kept as is.
Returns
~~~~~~~
PointList of the subsampled point cloud.
"""
if ptsnb is None:
return points
else:
nbPoints = len(points)
subset = sample(range(nbPoints), ptsnb)
return points.subset(subset)
[docs]
def filter_points(points, densityfilterratio=0.05, densityradius=None, k=16):
"""
Filter a point cloud using density estimation.
Parameters
~~~~~~~~~~
points: PointList
PointList of the point cloud, usualy obtained with :func:`load_points`.
densityfilterratio: float
Density ratio to use to filter the point cloud using density estimation.
Default to 0.05.
densityradius: float
Value of the radius to use to estimate the density. Default is None, which means that the
value is calculated as 1/100 of the max-min of the point cloud in the z direction.
k: int
Number of neighbors to use to estimate the density. Default to 16.
Returns
~~~~~~~
PointList of the filtered point cloud.
"""
# Determine default value of radius for density estimation
if densityradius is None:
mini, maxi = points.getZMinAndMaxIndex()
zdist = points[maxi].z - points[mini].z
densityradius = zdist / 100.0
# Connect all points to k closest neighbors
kclosests = k_closest_points_from_ann(points, k, True)
kclosests = connect_all_connex_components(points, kclosests, True)
# Determine the neighbors in a radius
rnbgs = r_neighborhoods(points, kclosests, densityradius)
# Estimate densities
densities = densities_from_r_neighborhood(rnbgs, densityradius)
# Determine value of density under which we filter
mind, maxd = densities.getMinAndMax(True)
densitythreshold = mind + (maxd - mind) * densityfilterratio
# Determine index of points to filter
nbPoints = len(points)
subset = [i for i in range(nbPoints) if densities[i] < densitythreshold]
# Return result
return points.opposite_subset(subset)
[docs]
def find_root(points):
"""
Find the root of the point cloud.
Notes
~~~~~
We define the center of the point cloud as (x_c, y_c, z_c). Then the root is defined
as the closest point from the cloud to the point (x_c, y_c, z_min)
Parameters
~~~~~~~~~~
points: PointList
PointList of the point cloud, usualy obtained with :func:`load_points`.
Returns
~~~~~~~
Tuple of the root node id, the minimum z value and the maximum z value.
"""
center = points.getCenter()
pminid, pmaxid = points.getZMinAndMaxIndex()
zmin = points[pminid].z
zmax = points[pmaxid].z
initp = center
initp.z = zmin
return points.findClosest(initp)[0], zmin, zmax
[docs]
def skeleton(
points,
skel_func=xu_method,
xu_binratio=50,
xu_neighbors=20,
gc_min_growth_length=None,
gc_max_growth_length=None,
gc_radius_func=None,
sc_growth_length=None,
sc_kill_distance_ratio=0.9,
sc_presence_distance_ratio=2.5,
sc_min_nb_pt_per_bud=5,
asc_min_growth_length=None,
asc_max_growth_length=None,
asc_radius_func=None,
asc_kill_distance_ratio=0.9,
asc_presence_distance_ratio=2.5,
asc_min_nb_pt_per_bud=5,
livny_contraction_nb=3,
livny_filtering_nb=5,
livny_min_edge_ratio=0.15,
):
"""
Topology reconstruction of a point cloud.
Parameters
~~~~~~~~~~
points: PointList
PointList of the point cloud used to reconstruct topology.
skel_func: function
Reconstruction algorithm to use. Default to `xu_method`.
Other possible algorithms are:
- `graphcolonization_method`: Graph Colonization Algorithm
- `spacecolonization_method`: Space colonization reconstruction
- `adaptivespacecolonization_method`: Adaptive Space Colonization Algorithm
- `livny_method_mtg`: Livny algorithm
xu_binratio: float
Value of the bin ratio to use to estimate the topology.
Only used for XU method. Default to 50.
xu_neighbors: int
Number of neighbors to use to estimate the topology.
Only used for XU method. Default to 20.
gc_min_growth_length: float
Minimum growth length for the graph colonization algorithm.
Only used for Graph Colonization algorithm. Default to zdist / 200.
gc_max_growth_length: float
Maximum growth length for the graph colonization algorithm.
Only used for Graph Colonization algorithm. Default to zdist / 50.
gc_radius_func: function
Density - bin length Relationship function for the graph colonization algorithm.
Only used for Graph Colonization algorithm. Default to a NurbsCurve2D function with custom parameters.
sc_growth_length: float
Growth length for the space colonization algorithm.
Only used for Space Colonization algorithm. Default to zdist / 50.
sc_kill_distance_ratio: float
Kill distance ratio for the space colonization algorithm.
Only used for Space Colonization algorithm. Default to 0.9.
sc_presence_distance_ratio: float
Perception distance ratio for the space colonization algorithm.
Only used for Space Colonization algorithm. Default to 2.5.
sc_min_nb_pt_per_bud: int
Minimum number of points per bud for the space colonization algorithm.
Only used for Space Colonization algorithm. Default to 5.
asc_max_growth_length: float
Maximum growth length for the adaptive space colonization algorithm.
Only used for Adaptive Space Colonization algorithm. Default to zdist / 50.
asc_min_growth_length: float
Minimum growth length for the adaptive space colonization algorithm.
Only used for Adaptive Space Colonization algorithm. Default to zdist / 100.
asc_radius_func: function
Density - Radius Relationship function for the adaptive space colonization algorithm.
Only used for Adaptive Space Colonization algorithm. Default to a NurbsCurve2D function with custom parameters.
asc_kill_distance_ratio: float
Kill distance ratio for the adaptive space colonization algorithm.
Only used for Adaptive Space Colonization algorithm. Default to 0.9.
asc_presence_distance_ratio: float
Perception distance ratio for the adaptive space colonization algorithm.
Only used for Adaptive Space Colonization algorithm. Default to 2.5.
asc_min_nb_pt_per_bud: int
Minimum number of points per bud for the adaptive space colonization algorithm.
Only used for Adaptive Space Colonization algorithm. Default to 5.
livny_contraction_nb: int
Number of contraction steps for the livny algorithm.
Only used for Livny algorithm. Default to 3.
livny_filtering_nb: int
Number of filtering steps for the livny algorithm.
Only used for Livny algorithm. Default to 5.
livny_min_edge_ratio: int
Ratio of the minimum edge size for the livny algorithm.
Only used for Livny algorithm. Default to 0.15.
Returns
~~~~~~~
openalea.mtg.MTG object representing the topology of the point cloud.
"""
root, zmin, zmax = find_root(points)
mtg = initialize_mtg(root)
zdist = zmax - zmin
vtx = list(mtg.vertices(mtg.max_scale()))
startfrom = vtx[0]
def call_xu():
binlength = zdist / xu_binratio
return skel_func(mtg, startfrom, points, binlength, xu_neighbors)
def call_gc():
densities = mtg.pointinfo.densities
max_length = (
asc_max_growth_length if gc_max_growth_length is not None else zdist / 50
)
min_length = (
asc_min_growth_length if gc_min_growth_length is not None else zdist / 200
)
if gc_radius_func is None:
radius_func = NurbsCurve2D(
[(0, 0, 1), (0.3, 0.3, 1), (0.7, 0.7, 1), (1, 1, 1)]
)
else:
radius_func = gc_radius_func
return skel_func(
mtg,
startfrom,
points,
densities,
min_length,
max_length,
QuantisedFunction(radius_func.__deepcopy__({})),
)
def call_sc():
length = sc_growth_length if sc_growth_length is not None else zdist / 50
return skel_func(
mtg,
startfrom,
points,
length,
sc_kill_distance_ratio,
sc_presence_distance_ratio,
sc_min_nb_pt_per_bud,
)
def call_asc():
densities = mtg.pointinfo.densities
max_length = (
asc_max_growth_length if asc_max_growth_length is not None else zdist / 50
)
min_length = (
asc_min_growth_length if asc_min_growth_length is not None else zdist / 100
)
if asc_radius_func is None:
radius_func = NurbsCurve2D(
[(0, 0, 1), (0.3, 0.3, 1), (0.7, 0.7, 1), (1, 1, 1)]
)
else:
radius_func = asc_radius_func
return skel_func(
mtg,
startfrom,
points,
densities,
min_length,
max_length,
QuantisedFunction(radius_func.__deepcopy__({})),
asc_kill_distance_ratio,
asc_presence_distance_ratio,
asc_min_nb_pt_per_bud,
)
def call_livny():
return skel_func(
mtg,
startfrom,
points,
livny_contraction_nb,
livny_filtering_nb,
livny_min_edge_ratio,
)
dispatch = {
xu_method: call_xu,
graphcolonization_method: call_gc,
spacecolonization_method: call_sc,
adaptivespacecolonization_method: call_asc,
livny_method_mtg: call_livny,
}
try:
return dispatch[skel_func]()
except KeyError:
raise ValueError(f"Unsupported skeleton function: {skel_func}")
