CloudCompare——计算点云的KD树并可视化

1.功能概述

  使用Tools>Sand box (research)>compute kd-tree访问此工具。
  该功能可以用于论文写作中的KD树绘图。

2.完整操作

3.算法源码

static double s_kdTreeMaxErrorPerCell = 0.1;
void MainWindow::doActionComputeKdTree()
{
	ccGenericPointCloud* cloud = nullptr;

	if ( haveOneSelection() )
	{
		ccHObject* ent = m_selectedEntities.back();
		bool lockedVertices;
		cloud = ccHObjectCaster::ToGenericPointCloud(ent,&lockedVertices);
		if (lockedVertices)
		{
			ccUtils::DisplayLockedVerticesWarning(ent->getName(),true);
			return;
		}
	}

	if (!cloud)
	{
		ccLog::Error("Selected one and only one point cloud or mesh!");
		return;
	}

	bool ok;
	s_kdTreeMaxErrorPerCell = QInputDialog::getDouble(this, "Compute Kd-tree", "Max error per leaf cell:", s_kdTreeMaxErrorPerCell, 1.0e-6, 1.0e6, 6, &ok);
	if (!ok)
		return;

	ccProgressDialog pDlg(true, this);

	//computation
	QElapsedTimer eTimer;
	eTimer.start();
	ccKdTree* kdtree = new ccKdTree(cloud);

	if (kdtree->build(s_kdTreeMaxErrorPerCell, CCLib::DistanceComputationTools::MAX_DIST_95_PERCENT, 4, 1000, &pDlg))
	{
		qint64 elapsedTime_ms = eTimer.elapsed();

		ccConsole::Print("[doActionComputeKdTree] Timing: %2.3f s",static_cast<double>(elapsedTime_ms)/1.0e3);
		cloud->setEnabled(true); //for mesh vertices!
		cloud->addChild(kdtree);
		kdtree->setDisplay(cloud->getDisplay());
		kdtree->setVisible(true);
		kdtree->prepareDisplayForRefresh();
#ifdef QT_DEBUG
		kdtree->convertCellIndexToSF();
#else
		kdtree->convertCellIndexToRandomColor();
#endif

		addToDB(kdtree);

		refreshAll();
		updateUI();
	}
	else
	{
		ccLog::Error("An error occurred!");
		delete kdtree;
		kdtree = nullptr;
	}
}

4.相关代码

[1] C++代码： PCL KD树的使用

[2] python代码： Open3D KDTree的使用

[3] matlab代码：matlab KD树的使用

[4] Open3D (C++) KD tree的使用