PCL聚类 之K-Means

1. K-Means 聚类原理

目标：把点云分成 KKK 个簇，使簇内点尽量接近簇中心，簇间尽量远。

假设点云：

2、数学原理

优化目标函数

算法步骤

3、Matlab

clc; clear;

% 生成示例点云
pts1 = randn(50,3) + [1 1 1];
pts2 = randn(50,3) + [3 2 1];
pts3 = randn(50,3) + [2 3 2];
points = [pts1; pts2; pts3];

% 聚类
K = 3;
[idx, C] = kmeans(points, K);

% 可视化
figure; hold on; grid on; axis equal;
colors = lines(K);
for k = 1:K
    scatter3(points(idx==k,1), points(idx==k,2), points(idx==k,3), 50, colors(k,:), 'filled');
end
xlabel('X'); ylabel('Y'); zlabel('Z');
title('K-Means 点云聚类示例');