写出python代码实现用坡度滤波方法将原始点云分为地面点和非地面点并进行可视化

以下是Python代码实现用坡度滤波方法将原始点云分为地面点和非地面点并进行可视化的示例：

import numpy as np
import open3d as o3d

# 读取点云数据
pcd = o3d.io.read_point_cloud("point_cloud.pcd")

# 计算点云法向量
pcd.estimate_normals()

# 计算点云每个点的坡度
slopes = []
for i in range(len(pcd.points)):
    neighbors = pcd.tree.query_radius(pcd.points[i], r=0.1)
    if len(neighbors) < 5:
        slopes.append(0)
    else:
        normal = np.mean(np.asarray(pcd.normals)[neighbors], axis=0)
        slope = np.arccos(np.abs(np.dot(normal, [0, 0, 1])))
        slopes.append(slope)

# 将点云分为地面点和非地面点
ground_points = []
non_ground_points = []
for i in range(len(pcd.points)):
    if slopes[i] < np.radians(10):
        ground_points.append(pcd.points[i])
    else:
        non_ground_points.append(pcd.points[i])

# 可视化地面点和非地面点
ground_pcd = o3d.geometry.PointCloud()
ground_pcd.points = o3d.utility.Vector3dVector(ground_points)
ground_pcd.paint_uniform_color([0, 1, 0])  # 绿色
non_ground_pcd = o3d.geometry.PointCloud()
non_ground_pcd.points = o3d.utility.Vector3dVector(non_ground_points)
non_ground_pcd.paint_uniform_color([1, 0, 0])  # 红色
o3d.visualization.draw_geometries([ground_pcd, non_ground_pcd])

如果您有任何问题，请随时问我。