CODEWITHSUNDEEP
python3dmeshpoint-clouds
edd1
edd1

Reputation: 102

Memory overhead when running Pygalmesh on a point cloud

I'm using the Pygalmesh library. I'm working with this library on a point cloud, which can vary in the number of vertices, with some having 2000 or more vertices. However, I'm having a problem when executing the generate_mesh method. I'm executing all of this through Google Colab. When the method starts, everything goes well until a certain time. After a few minutes, the RAM memory starts to reach its limit and then Colab stops executing. Is there a way to optimize it? Or am I implementing it incorrectly?

My code:

import pygalmesh
import numpy as np

class Custom(pygalmesh.DomainBase):
    def __init__(self, points):
        super().__init__()
        self.points = np.array(points)

    def eval(self, x):
        distances = np.linalg.norm(self.points - np.array(x), axis=1)
        return min(distances) - 0.1

    def get_bounding_sphere_squared_radius(self):

        center = np.mean(self.points, axis=0)
        max_distance = np.max(np.linalg.norm(self.points - center, axis=1))
        return (max_distance + 0.1) ** 2

verts = [
     [254.338, 32.655, 37.4208],
     [254.342, 32.6546, 37.4204],
     [254.352, 32.6535, 37.4195],
     .....
 ]

custom_mesh = Custom(verts)
mesh = pygalmesh.generate_mesh(custom_mesh, max_cell_circumradius=0.2)

I hope to be able to execute the method without the machine finishing before the deadline. I can visualize the results for different amounts of points.

Upvotes: 0

Views: 31

Answers (0)

Related Questions