本教程代码开源:GitHub 欢迎fork
文章目录
前言
在本教程中,将介绍如何使用 KdTree 查找特定点或位置的 K 个最近邻,还将介绍如何查找用户指定的某个半径内的所有邻居(在这种情况下是随机的) 。
理论入门
kd 树或 k 维树是计算机科学中使用的一种数据结构,用于在具有 k 维的空间中组织一定数量的点。它是一个二叉搜索树,对其施加了其他约束。Kd 树对于范围和最近邻搜索非常有用。出于我们的目的,我们通常只会处理三维的点云,因此我们所有的 kd 树都是三维的。kd 树的每一层使用垂直于相应轴的超平面沿特定维度拆分所有子节点。在树的根部,所有子节点都将根据第一维进行拆分(即,如果第一维坐标小于根,它将在左子树中,如果大于根,则显然将在左子树中右子树)。树中的每一层都在下一个维度上进行划分,一旦所有其他维度都用尽,则返回到第一个维度。构建 kd 树的最有效方法是使用像 Quick Sort 那样的分区方法,将中点放在根处,将一维值较小的所有内容放在左侧,右侧较大。然后在左子树和右子树上重复此过程,直到要分区的最后一棵树仅由一个元素组成。
来自[*]:
这是一个二维 kd 树的例子:
这是最近邻搜索如何工作的演示。
参考:https://pcl.readthedocs.io/projects/tutorials/en/latest/kdtree_search.html#kdtree-search
pclpy代码
import pclpy
from pclpy import pcl
import numpy as np
if __name__ == '__main__':
# 生成点云数据
cloud_size = 100
a = np.random.ranf(cloud_size * 3).reshape(-1, 3) * 1024
cloud = pcl.PointCloud.PointXYZ.from_array(a)
kdtree = pcl.kdtree.KdTreeFLANN.PointXYZ()
kdtree.setInputCloud(cloud)
searchPoint = pcl.point_types.PointXYZ()
searchPoint.x = np.random.ranf(1) * 1024
searchPoint.y = np.random.ranf(1) * 1024
searchPoint.z = np.random.ranf(1) * 1024
# k最近邻搜索
k = 8
pointIdxNKNSearch = pclpy.pcl.vectors.Int([0] * k)
pointNKNSquaredDistance = pclpy.pcl.vectors.Float([0] * k)
print('K nearest neighbor search at (', searchPoint.x,
'', searchPoint.y,
'', searchPoint.z,
') with k =', k, '\n')
if kdtree.nearestKSearch(searchPoint, k, pointIdxNKNSearch, pointNKNSquaredDistance) > 0:
for i in range(len(pointIdxNKNSearch)):
print(" ", cloud.x[pointIdxNKNSearch[i]],
" ", cloud.y[pointIdxNKNSearch[i]],
" ", cloud.z[pointIdxNKNSearch[i]],
" (squared distance: ", pointNKNSquaredDistance[i], ")", "\n")
# 使用半径最近邻搜索
pointIdxNKNSearch = pclpy.pcl.vectors.Int()
pointNKNSquaredDistance = pclpy.pcl.vectors.Float()
radius = np.random.ranf(1) * 256.0
print("Neighbors within radius search at (", searchPoint.x,
" ", searchPoint.y, " ", searchPoint.z, ") with radius=",
radius, '\n')
if kdtree.radiusSearch(searchPoint, radius, pointIdxNKNSearch, pointNKNSquaredDistance) > 0:
for i in range(len(pointIdxNKNSearch)):
print(" ", cloud.x[pointIdxNKNSearch[i]],
" ", cloud.y[pointIdxNKNSearch[i]],
" ", cloud.z[pointIdxNKNSearch[i]],
" (squared distance: ", pointNKNSquaredDistance[i], ")", "\n")
说明
以下代码首先利用Numpy使用随机数据创建和填充 PointCloud。
# 生成点云数据
cloud_size = 100
a = np.random.ranf(cloud_size * 3).reshape(-1, 3) * 1024
cloud = pcl.PointCloud.PointXYZ.from_array(a)
下一段代码创建我们的 kdtree 对象并将我们随机创建的云设置为输入。然后我们创建一个“searchPoint”,它被分配了随机坐标。
kdtree = pcl.kdtree.KdTreeFLANN.PointXYZ()
kdtree.setInputCloud(cloud)
searchPoint = pcl.point_types.PointXYZ()
searchPoint.x = np.random.ranf(1) * 1024
searchPoint.y = np.random.ranf(1) * 1024
searchPoint.z = np.random.ranf(1) * 1024
现在我们创建一个整数(并将其设置为 10)和两个向量,用于存储来自搜索的 K 个最近邻。
k = 8
pointIdxNKNSearch = pclpy.pcl.vectors.Int([0] * k)
pointNKNSquaredDistance = pclpy.pcl.vectors.Float([0] * k)
print('K nearest neighbor search at (', searchPoint.x,
'', searchPoint.y,
'', searchPoint.z,
') with k =', k, '\n')
假设我们的 KdTree 返回 0 个以上的最近邻,它然后打印出所有 10 个最近邻的位置到我们的随机“searchPoint”,这些位置已经存储在我们之前创建的向量中。
if kdtree.nearestKSearch(searchPoint, k, pointIdxNKNSearch, pointNKNSquaredDistance) > 0:
for i in range(len(pointIdxNKNSearch)):
print(" ", cloud.x[pointIdxNKNSearch[i]],
" ", cloud.y[pointIdxNKNSearch[i]],
" ", cloud.z[pointIdxNKNSearch[i]],
" (squared distance: ", pointNKNSquaredDistance[i], ")", "\n")
下面再演示一下使用半径最近邻搜索,在某个(随机生成的)半径内找到给定“searchPoint”的所有邻居。再次创建了 2 个向量来存储有关我们邻居的信息。
# 使用半径最近邻搜索
pointIdxNKNSearch = pclpy.pcl.vectors.Int()
pointNKNSquaredDistance = pclpy.pcl.vectors.Float()
radius = np.random.ranf(1) * 256.0
print("Neighbors within radius search at (", searchPoint.x,
" ", searchPoint.y, " ", searchPoint.z, ") with radius=",
radius, '\n')
同样,和以前一样,如果我们的 KdTree 在指定半径内返回 0 个以上的邻居,它会打印出这些点的坐标,这些点已经存储在我们的向量中。
if kdtree.radiusSearch(searchPoint, radius, pointIdxNKNSearch, pointNKNSquaredDistance) > 0:
for i in range(len(pointIdxNKNSearch)):
print(" ", cloud.x[pointIdxNKNSearch[i]],
" ", cloud.y[pointIdxNKNSearch[i]],
" ", cloud.z[pointIdxNKNSearch[i]],
" (squared distance: ", pointNKNSquaredDistance[i], ")", "\n")
运行
运行kdTreeDemo.py,即可
运行结果:
K nearest neighbor search at ( 737.6050415039062 750.3650512695312 411.2821960449219 ) with k = 8
753.9883 643.7369 456.1752 (squared distance: 13653.359375 )
828.60803 626.9191 501.09518 (squared distance: 31586.8125 )
760.72687 627.8939 539.5448 (squared distance: 31985.091796875 )
810.8796 972.1281 278.54584 (squared distance: 72166.953125 )
598.6487 507.64853 444.17035 (squared distance: 79301.8125 )
649.69885 946.6329 597.18005 (squared distance: 80806.5625 )
476.53268 646.1927 467.837 (squared distance: 82209.1015625 )
878.47424 922.55475 621.521 (squared distance: 93693.7734375 )
Neighbors within radius search at ( 737.6050415039062 750.3650512695312 411.2821960449219 ) with radius= [203.92877983]
753.9883 643.7369 456.1752 (squared distance: 13653.359375 )
828.60803 626.9191 501.09518 (squared distance: 31586.8125 )
760.72687 627.8939 539.5448 (squared distance: 31985.091796875 )
注意:由于我们的数据是随生成的,每次结果都不一样,甚至有时候可能KdTree 返回 0 最近邻,这时候就没有输出了。