File size: 3,614 Bytes
69f15ae
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
import trimesh
import json
import numpy as np

PI = np.pi

# 可以是标量(各向同性)或三元组(各向异性)
# 例如:1.0  或  [0.08, 0.4, 0.08]
MODEL_SCALE = 0.004

GRASP_OFFSET = [0., 0., 0.]   # 相对“顶部中心”的偏移
GRASP_ROT = [PI, 0.0, PI]             # 欧拉角 (rx, ry, rz), 弧度

def _to_scale_vec(scale):
    """将标量/数组统一成 3 维缩放向量 [sx, sy, sz]."""
    arr = np.asarray(scale, dtype=float)
    if arr.ndim == 0:
        return [float(arr)] * 3
    if arr.shape == (3,):
        return arr.tolist()
    raise ValueError(f"MODEL_SCALE 需为标量或长度为3的向量,收到形状: {arr.shape}")

def create_model_data(id="vase"):
    file_path = f"./{id}.glb"
    save_path = f"./model_data.json"

    # 1) 读模型并融合为单一 Trimesh
    with open(file_path, "rb") as f:
        loaded = trimesh.load(f, file_type="glb")
    mesh = loaded.to_mesh() if isinstance(loaded, trimesh.Scene) else loaded

    # 2) 应用缩放(标量或三元向量都可)
    scale_vec = _to_scale_vec(MODEL_SCALE)
    mesh.apply_scale(scale_vec)  # 支持 float 或 (3,) 向量

    # 3) OBB 信息
    obb = mesh.bounding_box_oriented
    center = obb.centroid
    ext = obb.extents
    diag = np.linalg.norm(ext)

    # 4) 计算“底部中心” -> 原点的平移
    rotation_matrix = obb.transform[:3, :3]
    bottom_center = center - rotation_matrix[:, 2] * (ext[2] / 2.0)
    mesh.apply_translation(-bottom_center)

    # 5) 更新 OBB
    obb = mesh.bounding_box_oriented
    center = obb.centroid
    ext = obb.extents
    rotation_matrix = obb.transform[:3, :3]

    # 6) 顶部中心(现在模型底部已对齐原点,Z 轴正向为“上”)
    top_center = center + rotation_matrix[:, 2] * (ext[2] / 2.0)

    # 7) 组合抓取位姿:先平移到顶部中心 -> 旋转 -> 偏移
    grasp_translation = trimesh.transformations.translation_matrix(top_center)
    grasp_rotation = trimesh.transformations.euler_matrix(*GRASP_ROT)
    grasp_offset = trimesh.transformations.translation_matrix(GRASP_OFFSET)
    grasp_transform = grasp_translation @ grasp_rotation @ grasp_offset

    # 8) 可视化
    scene = trimesh.Scene(mesh)
    grasp_sphere = trimesh.creation.icosphere(subdivisions=2, radius=diag * 0.05)
    grasp_sphere.apply_transform(grasp_transform)
    grasp_sphere.visual.vertex_colors = np.array([[1.0, 0.0, 0.0, 0.5]] * len(grasp_sphere.vertices))
    scene.add_geometry(grasp_sphere)

    grasp_axis = trimesh.creation.axis(origin_size=0.05, axis_length=diag * 0.2)
    grasp_axis.apply_transform(grasp_transform)
    scene.add_geometry(grasp_axis)

    # 9) 保存 JSON(注意 scale 为三元向量)
    data = {
        "center": center.tolist(),
        "extents": ext.tolist(),
        "scale": scale_vec,  
        "contact_points_pose": [grasp_transform.tolist()], 
        "transform_matrix": trimesh.transformations.identity_matrix().tolist(),
        "functional_matrix": [],
        "orientation_point": grasp_transform.tolist(),
        "contact_points_group": [],
        "contact_points_mask": [],
        "contact_points_discription": [],
        "target_point_discription": ["Grasp pose at the top edge center."],
        "functional_point_discription": [""],
        "orientation_point_discription": ["Grasp pose with adjustable offset and rotation."]
    }
    with open(save_path, "w", encoding="utf-8") as f:
        json.dump(data, f, indent=4, ensure_ascii=False, separators=(",", ": "))

    # 10) 显示场景
    scene.show()

if __name__ == "__main__":
    create_model_data("vase")