I am trying to find the Gaussian Mixture Model parameters of each colored cluster in the pointcloud shown below. I understand I can print out the GMM means and covariances of each cluster in the pointcloud, but when I visualize it, the clusters each have a unique color.
How do I find the GMM parameters corresponding to a certain color of a cluster from the pointcloud shown below? Here is my Python code:
import numpy as np
from sklearn.mixture import GaussianMixture
import open3d as o3d
import matplotlib.pyplot as plt
import pdb
def load_point_cloud(file_path):
pc = None
pcd = o3d.io.read_point_cloud(file_path)
pcd_points = np.asarray(pcd.points)
pcd_colors = np.asarray(pcd.colors)
pc = np.concatenate([pcd_points, pcd_colors], axis=1)
return pcd_points, pcd_colors, pc
def coords_to_pcl(coordarray):
xyz = coordarray[0:None, 0:3]
rgb = coordarray[0:None, 3:None]
pcd = o3d.geometry.PointCloud()
pcd.points = o3d.utility.Vector3dVector(xyz)
pcd.colors = o3d.utility.Vector3dVector(rgb)
return pcd
coord = o3d.geometry.TriangleMesh().create_coordinate_frame(size=0.25)
pc_xyz,_,_ = load_point_cloud("out.ply") #returns numpy array
gmm = GaussianMixture(n_components=8, random_state=0).fit(pc_xyz) #Estimate model parameters with the EM algorithm
labels = gmm.predict(pc_xyz)
max_label = labels.max()
print(f"point cloud has {max_label + 1} clusters from GMM")
colors = plt.get_cmap("tab20")(labels / (max_label if max_label > 0 else 1))
colors[labels<0] = 0
pcd = o3d.geometry.PointCloud()
pcd.points = o3d.utility.Vector3dVector(pc_xyz)
pcd.colors = o3d.utility.Vector3dVector(colors[:,:3])
o3d.visualization.draw_geometries([pcd, coord])
