DFJointSegmentator component#
Inputs:
|
The cluster point clouds from normal segmentation. |
|
The assembly to extract the joints from |
|
From 0 to 1, it’s the sin value. By default 0.1. The closer to 0 the less permissive and viceversa to 1. |
|
From 0 to infinite. By default 0.1. The closer to 0 the less permissive your point. |
|
The maximum correspondence distance for the local ICP registration that re-aligns the joints for per-face analysis of each joint. The default value is 0.005 |
|
The maximum distance for a joint segment to be considered as valid for the reference we are comparing to. The default value is 0.05 (5 cm). |
Outputs:
o_joint_face_segments |
The individual faces of the extracted joints. |
o_joint_segments |
The extracted joints. |
o_reference_point_clouds |
The reference point clouds of the joints. They are generated from the |
Code:
"""Extracts the joints from a point cloud."""
#! python3
import System
import math
import Rhino
import ghpythonlib.treehelpers
from diffCheck import diffcheck_bindings
from diffCheck import df_cvt_bindings as df_cvt
from ghpythonlib.componentbase import executingcomponent as component
ABSTOL = Rhino.RhinoDoc.ActiveDoc.ModelAbsoluteTolerance
class DFJointSegmentator(component):
def __init__(self):
super(DFJointSegmentator, self).__init__()
def RunScript(self,
i_clusters: System.Collections.Generic.List[Rhino.Geometry.PointCloud],
i_assembly,
i_angle_threshold: float,
i_distance_threshold: float,
i_correspondence_distance: float,
i_joint_displacement_tolerance: float):
if i_clusters is None or i_assembly is None:
return None
if i_angle_threshold is None:
i_angle_threshold = 0.1
if i_distance_threshold is None:
i_distance_threshold = 0.1
if i_correspondence_distance is None:
i_correspondence_distance = 0.005
if i_joint_displacement_tolerance is None:
i_joint_displacement_tolerance =0.05
if len(i_clusters) == 0:
raise ValueError("No clusters given.")
if not isinstance(i_clusters[0], Rhino.Geometry.PointCloud):
raise ValueError("The input clusters must be PointClouds.")
# get number of joints
n_joints = i_assembly.total_number_joints
# prepping the reference meshes
df_joints = [[] for _ in range(n_joints)]
rh_joints = [[] for _ in range(n_joints)]
for joint in i_assembly.all_joints:
for face in joint.faces:
face = face.to_mesh()
face.Subdivide()
face.Faces.ConvertQuadsToTriangles()
rh_joints[joint.id].append(face)
df_joints[joint.id].append(df_cvt.cvt_rhmesh_2_dfmesh(face))
o_reference_point_clouds = []
o_joint_faces_segments = []
df_cloud_clusters = [df_cvt.cvt_rhcloud_2_dfcloud(cluster) for cluster in i_clusters]
df_joint_clouds = []
o_joint_segments = []
# compute the center of the joints
rh_joint_centers = []
for rh_joint in rh_joints:
vertices = []
for face in rh_joint:
for vertice in face.Vertices:
vertices.append(Rhino.Geometry.Point3d(vertice.X, vertice.Y, vertice.Z))
joint_center = Rhino.Geometry.BoundingBox(vertices).Center
rh_joint_centers.append([joint_center.X, joint_center.Y, joint_center.Z])
# for each joint, find the corresponding faces, store them as such but also merge them, generate a reference point cloud, and register the merged clusters to the reference point cloud
for i, df_joint in enumerate(df_joints):
rh_joint_faces_segments = []
reference_joint_center = rh_joint_centers[i]
# create the reference point cloud
ref_df_joint_cloud = diffcheck_bindings.dfb_geometry.DFPointCloud()
for face in df_joint:
ref_face_cloud = face.sample_points_uniformly(1000)
ref_df_joint_cloud.add_points(ref_face_cloud)
o_reference_point_clouds.append(df_cvt.cvt_dfcloud_2_rhcloud(ref_df_joint_cloud))
# find the corresponding clusters and merge them
df_joint_cloud = diffcheck_bindings.dfb_geometry.DFPointCloud()
df_joint_face_segments = diffcheck_bindings.dfb_segmentation.DFSegmentation.associate_clusters(False, df_joint, df_cloud_clusters, i_angle_threshold, i_distance_threshold)
for df_joint_face_segment in df_joint_face_segments:
df_joint_cloud.add_points(df_joint_face_segment)
# get the center of the segment
if len(df_joint_cloud.points)>0:
df_cloud_bb_points = df_joint_cloud.get_tight_bounding_box()
x, y, z = 0, 0, 0
for i in range(len(df_cloud_bb_points)):
x += df_cloud_bb_points[i][0]
y += df_cloud_bb_points[i][1]
z += df_cloud_bb_points[i][2]
x = x/8 # because a bb has 8 corners
y = y/8
z = z/8
segment_center = [x, y, z]
segment_dist_to_ref = math.sqrt(math.pow(segment_center[0]-reference_joint_center[0], 2)
+ math.pow(segment_center[1]-reference_joint_center[1], 2)
+ math.pow(segment_center[2]-reference_joint_center[2], 2))
if segment_dist_to_ref > i_joint_displacement_tolerance:
rh_joint_cloud = df_cvt.cvt_dfcloud_2_rhcloud(df_joint_cloud)
rh_joint_cloud.SetUserString("df_sanity_scan_check", "1")
o_joint_segments.append(rh_joint_cloud)
else:
rh_joint_cloud = df_cvt.cvt_dfcloud_2_rhcloud(df_joint_cloud)
rh_joint_cloud.SetUserString("df_sanity_scan_check", "0")
o_joint_segments.append(rh_joint_cloud)
else:
rh_joint_cloud = df_cvt.cvt_dfcloud_2_rhcloud(df_joint_cloud)
rh_joint_cloud.SetUserString("df_sanity_scan_check", "2")
o_joint_segments.append(rh_joint_cloud)
# register the joint faces to the reference point cloud
transform = diffcheck_bindings.dfb_registrations.DFRefinedRegistration.O3DICP(df_joint_cloud, ref_df_joint_cloud, max_correspondence_distance = i_correspondence_distance)
for df_joint_face_segment in df_joint_face_segments:
df_joint_face_segment.apply_transformation(transform)
rh_joint_faces_segments.append(df_cvt.cvt_dfcloud_2_rhcloud(df_joint_face_segment))
df_joint_clouds.append(df_joint_cloud)
o_joint_faces_segments.append(rh_joint_faces_segments)
for rh_joint_faces, rh_joint in zip(o_joint_faces_segments, o_joint_segments):
for joint_face in rh_joint_faces:
joint_face.SetUserString("df_sanity_scan_check", rh_joint.GetUserString("df_sanity_scan_check"))
o_gh_tree_joint_faces_segments = ghpythonlib.treehelpers.list_to_tree(o_joint_faces_segments)
return o_gh_tree_joint_faces_segments, o_joint_segments, o_reference_point_clouds