点云处理：ICP配准

import open3d as o3d
import numpy as np

#读取电脑中的 ply 点云文件
source = o3d.io.read_point_cloud("/home/points/3.pcd")  #source 为需要配准的点云
target = o3d.io.read_point_cloud("/home/points/1.pcd")  #target 为目标点云

#为两个点云上上不同的颜色
source.paint_uniform_color([0.4, 0.6, 0.0])   #source 为黄色
target.paint_uniform_color([0.3, 0.3, 0.4])   #target 为蓝色

source = source.voxel_down_sample(voxel_size=0.05)
target = target.voxel_down_sample(voxel_size=0.05)

# source.remove_statistical_outlier(nb_neighbors=50,std_ratio=5)
# target.remove_statistical_outlier(nb_neighbors=50,std_ratio=5)

threshold = 1.5  #移动范围的阀值
trans_init = np.asarray([[1,0,0,0],   # 4x4 identity matrix，这是一个转换矩阵，
                         [0,1,0,0],   # 象征着没有任何位移，没有任何旋转，我们输入
                         [0,0,1,0],   # 这个矩阵为初始变换
                         [0,0,0,1]])

reg_p2p = o3d.pipelines.registration.registration_icp(source, target, threshold, trans_init, o3d.pipelines.registration.TransformationEstimationPointToPoint())
source.transform(reg_p2p.transformation)

# #创建一个 o3d.visualizer class
vis = o3d.visualization.Visualizer()
vis.create_window(window_name ='apple', width=1920, height=1080, left=0, top=0)

render_option = vis.get_render_option()
render_option.point_size = 1
render_option.background_color = np.asarray([0.1, 0.1, 0.2])

#将两个点云放入visualizer
vis.add_geometry(source)
vis.add_geometry(target)

vis.run()