sam-3d-objects/visualize_splat_static.py at main · Cooper-903/sam-3d-objects · GitHub

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
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
#!/usr/bin/env python3
"""
使用 PyTorch3D 渲染 splat.ply 点云为静态图像
"""

import torch
import matplotlib.pyplot as plt
from pytorch3d.io import IO
from pytorch3d.structures import Pointclouds
from pytorch3d.renderer import (
    PointsRasterizationSettings,
    PointsRenderer,
    PointsRasterizer,
    AlphaCompositor,
    PerspectiveCameras,
)
import numpy as np

def render_pointcloud_views(ply_path, output_dir="rendered_views", device='cuda', image_size=1024):
    """
    从多个角度渲染点云并保存图像

    Args:
        ply_path: PLY 文件路径
        output_dir: 输出目录
        image_size: 图像尺寸
    """
    import os
    os.makedirs(output_dir, exist_ok=True)

    print(f"正在加载 {ply_path}...")
    io = IO()
    pointcloud = io.load_pointcloud(ply_path, device=device)
    print(f"✓ 成功加载点云: {len(pointcloud.points_packed())} 个点")

    # 获取点云数据
    points = pointcloud.points_packed()
    features = pointcloud.features_packed()

    # 处理颜色
    if features is not None and features.shape[1] >= 3:
        colors = features[:, :3]
        if colors.max() > 1.0:
            colors = colors / 255.0
    else:
        # 使用 z 坐标着色
        z_coords = points[:, 2]
        z_normalized = (z_coords - z_coords.min()) / (z_coords.max() - z_coords.min() + 1e-8)
        colors = plt.cm.viridis(z_normalized.cpu().numpy())[:, :3]
        colors = torch.from_numpy(colors).float().to(device)

    pointcloud = Pointclouds(points=[points], features=[colors])

    # 计算点云的中心和范围
    center = points.mean(dim=0)
    scale = (points.max(dim=0)[0] - points.min(dim=0)[0]).max()
    distance = scale * 2.5

    # 设置多个视角
    views = [
        {"elevation": 10, "azimuth": 0, "name": "front"},
        {"elevation": 10, "azimuth": 90, "name": "right"},
        {"elevation": 10, "azimuth": 180, "name": "back"},
        {"elevation": 10, "azimuth": 270, "name": "left"},
        {"elevation": 45, "azimuth": 45, "name": "top_right"},
        {"elevation": -20, "azimuth": 45, "name": "bottom_right"},
    ]

    # 设置渲染器
    raster_settings = PointsRasterizationSettings(
        image_size=image_size,
        radius=0.003,
        points_per_pixel=10,
    )

    for view in views:
        print(f"正在渲染视角: {view['name']}...")

        # 计算相机位置
        elevation_rad = np.radians(view["elevation"])
        azimuth_rad = np.radians(view["azimuth"])

        x = distance * np.cos(elevation_rad) * np.sin(azimuth_rad)
        y = distance * np.sin(elevation_rad)
        z = distance * np.cos(elevation_rad) * np.cos(azimuth_rad)

        camera_position = torch.tensor([[x, y, z]], device=device) + center.unsqueeze(0)

        # 创建相机
        cameras = PerspectiveCameras(
            device=device,
            R=torch.eye(3, device=device).unsqueeze(0),
            T=-camera_position,
            focal_length=torch.tensor([[image_size * 1.2]], device=device),
            image_size=torch.tensor([[image_size, image_size]], device=device),
        )

        # 创建渲染器
        rasterizer = PointsRasterizer(
            cameras=cameras,
            raster_settings=raster_settings
        )

        renderer = PointsRenderer(
            rasterizer=rasterizer,
            compositor=AlphaCompositor()
        )

        # 渲染
        images = renderer(pointcloud)

        # 转换为 numpy 并保存
        image = images[0, ..., :3].cpu().numpy()
        image = np.clip(image, 0, 1)

        output_path = os.path.join(output_dir, f"splat_{view['name']}.png")
        plt.figure(figsize=(10, 10))
        plt.imshow(image)
        plt.axis('off')
        plt.title(f'Splat Point Cloud - {view["name"]} View')
        plt.tight_layout()
        plt.savefig(output_path, dpi=150, bbox_inches='tight')
        plt.close()
        print(f"  ✓ 已保存到: {output_path}")

    print(f"\n✓ 所有视角已渲染完成，保存在 {output_dir}/ 目录")

if __name__ == "__main__":
    import sys

    device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
    print(f"使用设备: {device}")

    ply_path = "splat.ply"
    if len(sys.argv) > 1:
        ply_path = sys.argv[1]

    render_pointcloud_views(ply_path, device=device)