Open3D体素法插入球体缺陷报错问题求助

报错原因与修复 #

1. AttributeError: VoxelGrid无get_voxel_center()属性 #

Open3D的VoxelGrid类从未提供过get_voxel_center()方法，属于API误用。正确获取体素中心的方式是通过体素的grid_index、体素网格的origin和voxel_size手动计算：

import numpy as np

# 计算单个体素的中心
voxel_center = voxel_grid.origin + (np.array(voxel.grid_index) + np.array([0.5, 0.5, 0.5])) * voxel_grid.voxel_size

2. TypeError: Voxel对象无法与浮点值运算 #

selected_voxel是Voxel类实例，不能直接参与数值运算。需提取其grid_index属性（整数元组），转换为numpy数组后再计算：

# 错误示例：直接用Voxel对象运算
# selected_voxel + 0.5  # 触发TypeError

# 正确方式：提取grid_index并转换为数组
grid_index = np.array(selected_voxel.grid_index)
voxel_center = voxel_grid.origin + (grid_index + 0.5) * voxel_grid.voxel_size

完整实现流程 #

以下是将球体缺陷插入空心对象实体区域（避开空域与顶点）的可运行代码：

步骤1：导入依赖并加载模型 #

import open3d as o3d
import numpy as np

# 加载空心对象与球体缺陷模型
hollow_mesh = o3d.io.read_triangle_mesh("hollow_object.obj")
defect_mesh = o3d.io.read_triangle_mesh("sphere_defect.ply")

# 修复网格（可选，解决可能的拓扑问题）
hollow_mesh.remove_duplicated_vertices()
hollow_mesh.remove_non_manifold_edges()
defect_mesh.remove_duplicated_vertices()

步骤2：体素化并筛选目标体素 #

# 设置体素大小（根据模型尺寸调整）
voxel_size = 0.01

# 对空心对象采样点云（覆盖实体区域）
solid_pcd = hollow_mesh.sample_points_uniformly(number_of_points=100000)
# 体素化点云，得到实体区域的体素网格
voxel_grid_solid = o3d.geometry.VoxelGrid.create_from_point_cloud(solid_pcd, voxel_size=voxel_size)

# 筛选：避开顶点附近的体素
vertex_coords = np.asarray(hollow_mesh.vertices)
distance_threshold = voxel_size * 1.5  # 可调整阈值
selected_voxels = []

for voxel in voxel_grid_solid.get_voxels():
    # 计算体素中心
    grid_idx = np.array(voxel.grid_index)
    center = voxel_grid_solid.origin + (grid_idx + 0.5) * voxel_grid_solid.voxel_size
    # 计算到所有顶点的最小距离
    min_dist = np.min(np.linalg.norm(vertex_coords - center, axis=1))
    if min_dist > distance_threshold:
        selected_voxels.append(voxel)

步骤3：插入球体缺陷 #

# 缩放缺陷至适配体素大小
defect_scale = voxel_size * 2
defect_mesh.scale(defect_scale, center=defect_mesh.get_center())

# 合并缺陷到空心对象
final_mesh = hollow_mesh.clone()
for voxel in selected_voxels:
    grid_idx = np.array(voxel.grid_index)
    center = voxel_grid_solid.origin + (grid_idx + 0.5) * voxel_grid_solid.voxel_size
    # 复制并平移缺陷到体素中心
    defect_copy = defect_mesh.clone()
    defect_copy.translate(center - defect_copy.get_center())
    final_mesh += defect_copy

# 保存结果
o3d.io.write_triangle_mesh("hollow_with_defects.obj", final_mesh)

关于trimesh/VTK体素化失败的补充建议 #

• trimesh：若体素化失败，可先修复网格（mesh.fill_holes()、mesh.merge_vertices()），再用trimesh.voxel.VoxelMesh.from_mesh(mesh, voxel_size=xxx, method='ray')，ray方法能更准确检测实体内部区域。
• VTK：若体素化失败，需确保网格是闭合的，可使用vtk.vtkCleanPolyData清理网格，再用vtk.vtkVoxelModeller时明确设置SetSampleDimensions和SetModelBounds参数，保证体素覆盖整个实体区域。

内容的提问来源于stack exchange，提问作者Anshul Dhingra