Linux下使用open3d进行点云可视化（.bin文件）

整个场景可视化：

import numpy as np
import open3d as o3d

def read_kitti_bin_point_cloud(bin_file):
    # 加载.bin文件
    point_cloud_np = np.fromfile(bin_file, dtype=np.float32).reshape(-1, 4)
    # 仅使用X, Y, Z坐标，忽略反射率
    point_cloud_o3d = o3d.geometry.PointCloud()
    point_cloud_o3d.points = o3d.utility.Vector3dVector(point_cloud_np[:, :3])
    # 设置所有点为白色
    point_cloud_o3d.colors = o3d.utility.Vector3dVector(np.ones((len(point_cloud_np), 3)))
    return point_cloud_o3d

def visualize_point_cloud(point_cloud_o3d):
    # 创建一个可视化窗口
    vis = o3d.visualization.Visualizer()
    vis.create_window(window_name="KITTI Point Cloud", width=800, height=600)
    # 设置窗口的背景为黑色
    opt = vis.get_render_option()
    opt.point_size = 1.0
    opt.background_color = np.asarray([0, 0, 0])
    # 将点云数据加入可视化窗口
    vis.add_geometry(point_cloud_o3d)
    # 进行可视化
    vis.run()
    vis.destroy_window()

if __name__ == "__main__":
    bin_file = '/root/autodl-data/IA-SSD/data/kitti/training/velodyne/002147.bin'  # 更改为你的KITTI .bin文件路径
    point_cloud_o3d = read_kitti_bin_point_cloud(bin_file)
    visualize_point_cloud(point_cloud_o3d)
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

效果：

KITTI范围：point_cloud_range = [0, -40, -3, 70.4, 40, 1]

import numpy as np
import open3d as o3d

def read_kitti_bin_point_cloud(bin_file, point_cloud_range):
    # 加载.bin文件
    point_cloud_np = np.fromfile(bin_file, dtype=np.float32).reshape(-1, 4)
    # 应用点云范围筛选
    mask = (point_cloud_np[:, 0] >= point_cloud_range[0]) & (point_cloud_np[:, 0] <= point_cloud_range[3]) & \
           (point_cloud_np[:, 1] >= point_cloud_range[1]) & (point_cloud_np[:, 1] <= point_cloud_range[4]) & \
           (point_cloud_np[:, 2] >= point_cloud_range[2]) & (point_cloud_np[:, 2] <= point_cloud_range[5])
    point_cloud_np = point_cloud_np[mask][:, :3]
    # 使用X, Y, Z坐标创建Open3D点云对象
    point_cloud_o3d = o3d.geometry.PointCloud()
    point_cloud_o3d.points = o3d.utility.Vector3dVector(point_cloud_np)
    # 设置所有点为白色
    point_cloud_o3d.colors = o3d.utility.Vector3dVector(np.ones((len(point_cloud_np), 3)))
    return point_cloud_o3d

def visualize_point_cloud(point_cloud_o3d):
    # 创建一个可视化窗口
    vis = o3d.visualization.Visualizer()
    vis.create_window(window_name="KITTI Point Cloud", width=800, height=600)
    # 设置窗口的背景为黑色
    opt = vis.get_render_option()
    opt.point_size = 1.0
    opt.background_color = np.asarray([0, 0, 0])
    # 将点云数据加入可视化窗口
    vis.add_geometry(point_cloud_o3d)
    # 进行可视化
    vis.run()
    vis.destroy_window()

if __name__ == "__main__":
    bin_file = '/root/autodl-data/IA-SSD/data/kitti/training/velodyne/002147.bin'  # 更改为你的KITTI .bin文件路径
    # 定义点云显示范围
    point_cloud_range = [0, -40, -3, 70.4, 40, 1]
    point_cloud_o3d = read_kitti_bin_point_cloud(bin_file, point_cloud_range)
    visualize_point_cloud(point_cloud_o3d)

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

效果：