Hey there! I’ve spent a fair amount of time working with 3D model manipulation in Python, so let me break down the best tools and hands-on workflows for handling .obj and .stl files—covering everything from loading, interactive viewing, adding multiple objects, to editing mesh data.

PyVista is hands-down the easiest library for interactive 3D in Python. It wraps VTK under the hood but has a Pythonic API that’s way more approachable for everyday tasks.

Installation #

First, get it installed:

pip install pyvista

Loading .obj/.stl Files #

Loading models is a total one-liner—no messy parsing required:

import pyvista as pv

# Load an .obj file
mesh_obj = pv.read("your_model.obj")

# Load an .stl file
mesh_stl = pv.read("your_model.stl")

Interactive Viewing #

To view the model with full interactivity (zoom, pan, rotate, toggle wireframe, etc.), just call:

mesh_obj.plot()

You’ll get a dedicated window with a toolbar—hover over buttons to see their functions. It’s intuitive enough that you’ll be navigating the model in seconds.

Adding Multiple Objects to the Scene #

Want to combine multiple models or add primitive shapes? Use a Plotter object to build your scene:

plotter = pv.Plotter()

# Add first mesh with a red color and label it
plotter.add_mesh(mesh_obj, color="red", label="Original Object")

# Shift the STL mesh along the X-axis and add it
mesh_stl_shifted = mesh_stl.translate([10, 0, 0])
plotter.add_mesh(mesh_stl_shifted, color="blue", label="Shifted STL")

# Add a small sphere at the origin as a reference
sphere = pv.Sphere(radius=1, center=[0,0,0])
plotter.add_mesh(sphere, color="green", label="Reference Sphere")

# Add a legend and show the final scene
plotter.add_legend()
plotter.show()

Editing Operations #

PyVista has tons of built-in editing tools. Here are a few common ones:

• Cropping: Cut a section out of the mesh with a bounding box

  # Crop to a box from (-5,-5,-5) to (5,5,5)
  cropped_mesh = mesh_obj.clip_box(bounds=[-5,5,-5,5,-5,5])
  cropped_mesh.plot(title="Cropped Mesh")
  

• Smoothing: Reduce mesh roughness for cleaner visuals

  smoothed_mesh = mesh_obj.smooth(n_iter=100)
  smoothed_mesh.plot(title="Smoothed Mesh")
  

• Merging Meshes: Combine your model with another shape

  cube = pv.Cube(center=[5,0,0], x_length=4)
  combined_mesh = mesh_obj.merge(cube)
  combined_mesh.plot(title="Combined Mesh")
  

If you need low-level control over mesh data (like modifying vertices, faces, or processing point clouds), Open3D is your go-to. It also has solid interactive tools for inspecting edits.

Installation #

Install via pip:

pip install open3d

Loading .obj/.stl Files #

import open3d as o3d

# Load .obj file
mesh_obj = o3d.io.read_triangle_mesh("your_model.obj")
# Load .stl file
mesh_stl = o3d.io.read_triangle_mesh("your_model.stl")

# Fix common mesh issues (like non-manifold edges)
mesh_obj.remove_non_manifold_edges()

Interactive Viewing #

Open3D’s viewer lets you interact with the model and even select individual faces/vertices:

o3d.visualization.draw_geometries([mesh_obj])

Mouse controls: Left-click drag to rotate, right-click drag to pan, scroll to zoom.

Adding Multiple Objects #

Just pass a list of geometries to the viewer function:

# Shift the STL mesh along the Y-axis
mesh_stl.translate([0, 6, 0])

# Add a cylinder as a third object
cylinder = o3d.geometry.TriangleMesh.create_cylinder(radius=1, height=8)
cylinder.translate([6,0,0])

# Show all objects together
o3d.visualization.draw_geometries([mesh_obj, mesh_stl, cylinder])

Editing Operations #

• Mesh Simplification: Reduce the number of faces for lighter models

  # Simplify to 1000 triangles from the original count
  simplified_mesh = mesh_obj.simplify_quadric_decimation(target_number_of_triangles=1000)
  o3d.visualization.draw_geometries([simplified_mesh], window_name="Simplified Mesh")
  

• Manual Vertex Editing: Adjust specific vertex positions

  import numpy as np
  
  # Get vertex data as a numpy array
  vertices = np.asarray(mesh_obj.vertices)
  # Move the first 15 vertices up along the Z-axis
  vertices[:15, 2] += 4
  # Update the mesh with modified vertices
  mesh_obj.vertices = o3d.utility.Vector3dVector(vertices)
  o3d.visualization.draw_geometries([mesh_obj], window_name="Edited Vertices")
  

• Plane Cutting: Split the mesh with a custom plane

  # Create a plane at the origin with normal along the X-axis
  plane = o3d.geometry.Plane()
  plane.set_from_point_normal([0,0,0], [1,0,0])
  # Cut the mesh along the plane
  cut_mesh = mesh_obj.cut_plane(plane)
  o3d.visualization.draw_geometries([cut_mesh], window_name="Cut Mesh")
  

If you don’t need heavy interactivity and just want to embed a 3D view in a matplotlib figure, this combo works. Note: Matplotlib’s 3D controls are clunky compared to PyVista/Open3D, so it’s best for quick previews.

Installation #

You’ll need matplotlib and trimesh (to handle mesh file loading):

pip install matplotlib trimesh

Loading and Viewing #

import matplotlib.pyplot as plt
from mpl_toolkits.mplot3d import Axes3D
import trimesh

# Load the mesh
mesh = trimesh.load("your_model.obj")

# Set up the 3D plot
fig = plt.figure(figsize=(10,8))
ax = fig.add_subplot(111, projection='3d')

# Plot each face of the mesh
for face in mesh.faces:
    ax.plot3D(mesh.vertices[face, 0], mesh.vertices[face, 1], mesh.vertices[face, 2], color='gray', linewidth=0.5)

# Add axis labels
ax.set_xlabel('X Axis')
ax.set_ylabel('Y Axis')
ax.set_zlabel('Z Axis')
ax.set_title('Quick Mesh Preview')

plt.show()

You can rotate the plot with mouse drag, but zoom and pan are not as smooth as the other libraries.

内容的提问来源于stack exchange，提问作者Ray Lee