Open3D 分割点云案例

**

Open3D 分割案例：

**
RANSAC 总结：（参数：最少点个数MinPts，半径 R）
1：确定拟合模型
2：确定拟合该模型每次需要的点个数
3：随机选点计算模型参数
4：根据距离d计算inlier个数
5：重复随机，保留inlier个数最多的那个模型
参数：距离d，随机次数n，拟合模型

利用前面所学知识，实践分割案例：
分割如下点云，实现步骤：

RANSAC：

import open3d as o3d
import numpy as np
import matplotlib.pyplot as plt
# Open3D 0.16版本的 api

pcd = o3d.io.read_point_cloud("sample.ply")
# 计算法线
pcd.estimate_normals(search_param=o3d.geometry.KDTreeSearchParamHybrid(radius=0.1, max_nn=16), fast_normal_computation=True)
pcd.paint_uniform_color([0,0,1])
# #RANSAC
[a, b, c, d], inliers = pcd.segment_plane(distance_threshold=0.01,ransac_n=3,num_iterations=1000)
print(f"平面方程: {a:.2f}x + {b:.2f}y + {c:.2f}z + {d:.2f} = 0")
inlier_cloud = pcd.select_by_index(inliers)
outlier_cloud = pcd.select_by_index(inliers, invert=True)
inlier_cloud.paint_uniform_color([1.0, 0, 0])
outlier_cloud.paint_uniform_color([0.6, 0.6, 0.6])
o3d.visualization.draw_geometries([inlier_cloud, outlier_cloud])

DBSCAN：

import open3d as o3d
import numpy as np
import matplotlib.pyplot as plt
# Open3D 0.16版本的 api

pcd = o3d.io.read_point_cloud("sample.ply")
# 计算法线
pcd.estimate_normals(search_param=o3d.geometry.KDTreeSearchParamHybrid(radius=0.1, max_nn=16), fast_normal_computation=True)
pcd.paint_uniform_color([0,0,1])
# # DBSCAN
labels = np.array(pcd.cluster_dbscan(eps=0.05, min_points=10))
max_label = labels.max()
print(f"point cloud has {max_label + 1} clusters")

# 给不同类别赋予不同的颜色
colors = plt.get_cmap("tab20")(labels / (max_label if max_label > 0 else 1))
colors[labels < 0] = 0
pcd.colors = o3d.utility.Vector3dVector(colors[:, :3])
o3d.visualization.draw_geometries([pcd])

segment_models={}
segments={}
max_plane_idx=40
rest=pcd
d_threshold=0.01
for i in range(max_plane_idx):
    colors = plt.get_cmap("tab20")(i)
    segment_models[i], inliers = rest.segment_plane(distance_threshold=0.01,ransac_n=3,num_iterations=1000)
    segments[i]=rest.select_by_index(inliers)
    labels = np.array(segments[i].cluster_dbscan(eps=d_threshold*10, min_points=10))
    candidates=[len(np.where(labels==j)[0]) for j in np.unique(labels)]
    best_candidate=int(np.unique(labels)[np.where(candidates==np.max(candidates))[0]])
    print("the best candidate is: ", best_candidate)
    rest = rest.select_by_index(inliers, invert=True)+segments[i].select_by_index(list(np.where(labels!=best_candidate)[0]))
    segments[i]=segments[i].select_by_index(list(np.where(labels==best_candidate)[0]))
    segments[i].paint_uniform_color(list(colors[:3]))
    print("pass",i+1,"/",max_plane_idx,"done.")

    labels = np.array(rest.cluster_dbscan(eps=0.05, min_points=5))
    max_label = labels.max()
    print(f"point cloud has {max_label + 1} clusters")

    colors = plt.get_cmap("tab10")(labels / (max_label if max_label > 0 else 1))
    colors[labels < 0] = 0
    rest.colors = o3d.utility.Vector3dVector(colors[:, :3])

o3d.visualization.draw_geometries([segments[i] for i in range(max_plane_idx)]+[rest])

运行效果如下:

代码难点解读，请看我下一篇文章，链接：
代码解析链接https://blog.csdn.net/qq_52040083/article/details/134435780?spm=1001.2014.3001.5502