[CC]DgmOctree—执行Cell遍历和单元计算
DgmOctree类的executeFunctionForAllCellsAtLevel和executeFunctionForAllCellsStartingAtLevel方法通过回调函数octreeCellFunc,执行八叉树中每个单元的相关计算。
1 2 3 4 5 6 7 | unsigned DgmOctree::executeFunctionForAllCellsAtLevel( unsigned char level, octreeCellFunc func, void ** additionalParameters, bool multiThread /*=false*/ , GenericProgressCallback* progressCb /*=0*/ , const char * functionTitle /*=0*/ , int maxThreadCount /*=0*/ ) |
1 2 3 4 5 6 7 8 9 | unsigned DgmOctree::executeFunctionForAllCellsStartingAtLevel(unsigned char startingLevel, octreeCellFunc func, void ** additionalParameters, unsigned minNumberOfPointsPerCell, unsigned maxNumberOfPointsPerCell, bool multiThread /* = true*/ , GenericProgressCallback* progressCb /*=0*/ , const char * functionTitle /*=0*/ , int maxThreadCount /*=0*/ ) |
通过DgmOctree遍历获取每一个Cell单元似乎还是一件很复杂的事情啊!
1 2 | //! The coded octree structure cellsContainer m_thePointsAndTheirCellCodes; |
通过编码集合实现?m_thePointsAndTheirCellCodes根据CellCodes对所有的点进行了排序。注意,所以知道每个Cell中的起始点的Index就可以确定Cell中的其他点。
1 2 3 4 | const cellsContainer& pointsAndTheirCellCodes() const { return m_thePointsAndTheirCellCodes; } |
注意octreeCellFunc 回调函数的参数,第一个是const引用传递参数,值不可以修改。
1 | typedef bool (*octreeCellFunc)( const octreeCell& cell, void **, NormalizedProgress*); |
因此尝试了使用下面的代码,应该是可行的:
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 | if (!m_app) return ; const ccHObject::Container& selectedEntities = m_app->getSelectedEntities(); size_t selNum = selectedEntities.size(); if (selNum!=1) { m_app->dispToConsole( "Select only one cloud!" ,ccMainAppInterface::ERR_CONSOLE_MESSAGE); return ; } ccHObject* ent = selectedEntities[0]; assert (ent); if (!ent || !ent->isA(CC_TYPES::POINT_CLOUD)) { m_app->dispToConsole( "Select a real point cloud!" ,ccMainAppInterface::ERR_CONSOLE_MESSAGE); return ; } ccPointCloud* theCloud = static_cast <ccPointCloud*>(ent); if (!theCloud) return ; //输入参数,半径大小 float kernelRadius=1; unsigned count = theCloud->size(); bool hasNorms = theCloud->hasNormals(); CCVector3 bbMin, bbMax; theCloud->getBoundingBox(bbMin,bbMax); const CCVector3d& globalShift = theCloud->getGlobalShift(); double globalScale = theCloud->getGlobalScale(); //进度条 ccProgressDialog pDlg( true ,m_app->getMainWindow()); unsigned numberOfPoints = theCloud->size(); if (numberOfPoints < 5) return ; //构建八叉树 CCLib::DgmOctree* theOctree = NULL; if (!theOctree) { theOctree = new CCLib::DgmOctree(theCloud); if (theOctree->build(&pDlg) < 1) { delete theOctree; return ; } } int result = 0; QString sfName= "Eigen_Value" ; int sfIdx = -1; ccPointCloud* pc = 0; //注意先添加ScalarField,并设置为当前的 if (theCloud->isA(CC_TYPES::POINT_CLOUD)) { pc = static_cast <ccPointCloud*>(theCloud); sfIdx = pc->getScalarFieldIndexByName(qPrintable(sfName)); if (sfIdx < 0) sfIdx = pc->addScalarField(qPrintable(sfName)); if (sfIdx >= 0) pc->setCurrentInScalarField(sfIdx); else { ccConsole::Error(QString( "Failed to create scalar field on cloud '%1' (not enough memory?)" ).arg(pc->getName())); } } //开启ScalarField模式 theCloud->enableScalarField(); //给定的半径,寻找最佳的Level unsigned char level = theOctree->findBestLevelForAGivenNeighbourhoodSizeExtraction(kernelRadius); //parameters void * additionalParameters[1] = { static_cast < void *>(&kernelRadius) }; if (theOctree->executeFunctionForAllCellsAtLevel(level, &computeCellEigenValueAtLevel, additionalParameters, true , &pDlg, "Eigen value Computation" ) == 0) { //something went wrong result = -4; } //number of cells for this level unsigned cellCount = theOctree->getCellNumber(level); unsigned maxCellPopulation =theOctree->getmaxCellPopulation(level); //cell descriptor (initialize it with first cell/point) CCLib::DgmOctree::octreeCell cell(theOctree); if (!cell.points->reserve(maxCellPopulation)) //not enough memory return ; cell.level = level; cell.index = 0; CCLib::DgmOctree::cellIndexesContainer vec; try { vec.resize(cellCount); } catch ( const std::bad_alloc&) { //not enough memory return ; } //binary shift for cell code truncation unsigned char bitDec = GET_NDT_BIT_SHIFT(level); CCLib::DgmOctree::cellsContainer::const_iterator p =theOctree->pointsAndTheirCellCodes().begin(); CCLib::DgmOctree::OctreeCellCodeType predCode = (p->theCode >> bitDec)+1; //pred value must be different than the first element's for (unsigned i=0,j=0; i<theOctree->getNumberOfProjectedPoints(); ++i,++p) { CCLib::DgmOctree::OctreeCellCodeType currentCode = (p->theCode >> bitDec); if (predCode != currentCode) { vec[j++] = i; //存储索引 int n=cell.points->size(); if (n>=5) { CCVector3d Psum(0,0,0); for (unsigned i=0; i<n; ++i) { CCVector3 P; cell.points->getPoint(i,P); Psum.x += P.x; Psum.y += P.y; Psum.z += P.z; } ScalarType curv = NAN_VALUE; CCVector3 G( static_cast <PointCoordinateType>(Psum.x / n), static_cast <PointCoordinateType>(Psum.y / n), static_cast <PointCoordinateType>(Psum.z / n) ); double mXX = 0.0; double mYY = 0.0; double mZZ = 0.0; double mXY = 0.0; double mXZ = 0.0; double mYZ = 0.0; //for each point in the cell for (unsigned i=0; i<n; ++i) { CCVector3 CellP; cell.points->getPoint(i,CellP); CCVector3 P = CellP - G; mXX += static_cast < double >(P.x)*P.x; mYY += static_cast < double >(P.y)*P.y; mZZ += static_cast < double >(P.z)*P.z; mXY += static_cast < double >(P.x)*P.y; mXZ += static_cast < double >(P.x)*P.z; mYZ += static_cast < double >(P.y)*P.z; } //symmetry CCLib::SquareMatrixd covMat(3); covMat.m_values[0][0] = mXX/n; covMat.m_values[1][1] = mYY/n; covMat.m_values[2][2] = mZZ/n; covMat.m_values[1][0] = covMat.m_values[0][1] = mXY/n; covMat.m_values[2][0] = covMat.m_values[0][2] = mXZ/n; covMat.m_values[2][1] = covMat.m_values[1][2] = mYZ/n; CCLib::SquareMatrixd eigVectors; std::vector< double > eigValues; if (!Jacobi< double >::ComputeEigenValuesAndVectors(covMat, eigVectors, eigValues)) { //failure curv= NAN_VALUE; } //compute eigen value double e0 = eigValues[0]; double e1 = eigValues[1]; double e2 = eigValues[2]; double sum = fabs (e0+e1+e2); if (sum < ZERO_TOLERANCE) { curv= NAN_VALUE; } double eMin = std::min(std::min(e0,e1),e2); curv= static_cast <ScalarType>( fabs (eMin) / sum); for (unsigned i=0; i<n; ++i) { //current point index unsigned index = cell.points->getPointGlobalIndex(i); //current point index in neighbourhood (to compute curvature at the right position!) unsigned indexInNeighbourhood = 0; cell.points->setPointScalarValue(i,curv); } } //and we start a new cell开始新的Cell cell.index += cell.points->size(); cell.points->clear( false ); cell.truncatedCode = currentCode; /*cell.mean_=G; cell.cov_=covMat; CCVector3 eigvalues1(e0,e1,e2); cell.evals_=eigvalues1;*/ } cell.points->addPointIndex(p->theIndex); //can't fail (see above) predCode = currentCode; } if (result == 0) { if (pc && sfIdx >= 0) { //设置当前显示的ScalarField pc->setCurrentDisplayedScalarField(sfIdx); pc->showSF(sfIdx >= 0); pc->getCurrentInScalarField()->computeMinAndMax(); } theCloud->prepareDisplayForRefresh(); } else { ccConsole::Warning(QString( "Failed to apply processing to cloud '%1'" ).arg(theCloud->getName())); if (pc && sfIdx >= 0) { pc->deleteScalarField(sfIdx); sfIdx = -1; } } |
参考DgmOctree::getCellIndexes DgmOctree::getPointsInCellByCellIndex

1 //重要,获取每一八叉树层的Cell索引的集合 2 bool DgmOctreeNDT::getCellIndexes(unsigned char level, cellIndexesContainer& vec) const 3 { 4 try 5 { 6 vec.resize(m_cellCount[level]); 7 } 8 catch (const std::bad_alloc&) 9 { 10 //not enough memory 11 return false; 12 } 13 14 //binary shift for cell code truncation 15 unsigned char bitDec = GET_NDT_BIT_SHIFT(level); 16 17 cellsContainer::const_iterator p = m_thePointsAndTheirCellCodes.begin(); 18 19 OctreeCellCodeType predCode = (p->theCode >> bitDec)+1; //pred value must be different than the first element's 20 21 for (unsigned i=0,j=0; i<m_numberOfProjectedPoints; ++i,++p) 22 { 23 OctreeCellCodeType currentCode = (p->theCode >> bitDec); 24 25 if (predCode != currentCode) 26 vec[j++] = i; 27 28 predCode = currentCode; 29 } 30 31 return true; 32 }

1 bool DgmOctreeNDT::getPointsInCellByCellIndex( ReferenceCloud* cloud, 2 unsigned cellIndex, 3 unsigned char level, 4 bool clearOutputCloud/* = true*/) const 5 { 6 assert(cloud && cloud->getAssociatedCloud() == m_theAssociatedCloud); 7 8 //binary shift for cell code truncation 9 unsigned char bitDec = GET_NDT_BIT_SHIFT(level); 10 11 //we look for the first index in 'm_thePointsAndTheirCellCodes' corresponding to this cell 12 cellsContainer::const_iterator p = m_thePointsAndTheirCellCodes.begin()+cellIndex; 13 OctreeCellCodeType searchCode = (p->theCode >> bitDec); 14 15 if (clearOutputCloud) 16 cloud->clear(false); 17 18 //while the (partial) cell code matches this cell 19 while ((p != m_thePointsAndTheirCellCodes.end()) && ((p->theCode >> bitDec) == searchCode)) 20 { 21 if (!cloud->addPointIndex(p->theIndex)) 22 return false; 23 ++p; 24 } 25 26 return true; 27 }
涉及的类包括:近邻点类、近邻点查询结构体
另外:

1 bool qNDTRansacSD::computeCellEigenValueAtLevel(const CCNDTLib::DgmOctreeNDT::octreeCellNDT& cell, 2 void** additionalParameters,NormalizedProgress* nProgress/*=0*/) 3 { 4 //parameters 5 PointCoordinateType radius = *static_cast<PointCoordinateType*>(additionalParameters[0]); 6 7 structure for nearest neighbors search 8 int n=cell.points->size(); 9 CCVector3d Psum(0,0,0); 10 for (unsigned i=0; i<n; ++i) 11 { 12 CCVector3 P; 13 cell.points->getPoint(i,P); 14 Psum.x += P.x; 15 Psum.y += P.y; 16 Psum.z += P.z; 17 } 18 ScalarType curv = NAN_VALUE; 19 CCVector3 G(static_cast<PointCoordinateType>(Psum.x / n), 20 static_cast<PointCoordinateType>(Psum.y / n), 21 static_cast<PointCoordinateType>(Psum.z / n) ); 22 23 double mXX = 0.0; 24 double mYY = 0.0; 25 double mZZ = 0.0; 26 double mXY = 0.0; 27 double mXZ = 0.0; 28 double mYZ = 0.0; 29 //for each point in the cell 30 for (unsigned i=0; i<n; ++i) 31 { 32 CCVector3 CellP; 33 cell.points->getPoint(i,CellP); 34 CCVector3 P = CellP - G; 35 36 mXX += static_cast<double>(P.x)*P.x; 37 mYY += static_cast<double>(P.y)*P.y; 38 mZZ += static_cast<double>(P.z)*P.z; 39 mXY += static_cast<double>(P.x)*P.y; 40 mXZ += static_cast<double>(P.x)*P.z; 41 mYZ += static_cast<double>(P.y)*P.z; 42 } 43 //symmetry 44 CCLib::SquareMatrixd covMat(3); 45 covMat.m_values[0][0] = mXX/n; 46 covMat.m_values[1][1] = mYY/n; 47 covMat.m_values[2][2] = mZZ/n; 48 covMat.m_values[1][0] = covMat.m_values[0][1] = mXY/n; 49 covMat.m_values[2][0] = covMat.m_values[0][2] = mXZ/n; 50 covMat.m_values[2][1] = covMat.m_values[1][2] = mYZ/n; 51 52 CCLib::SquareMatrixd eigVectors; 53 std::vector<double> eigValues; 54 55 if (!Jacobi<double>::ComputeEigenValuesAndVectors(covMat, eigVectors, eigValues)) 56 { 57 //failure 58 curv= NAN_VALUE; 59 } 60 61 //compute curvature as the rate of change of the surface 62 double e0 = eigValues[0]; 63 double e1 = eigValues[1]; 64 double e2 = eigValues[2]; 65 double sum = fabs(e0+e1+e2); 66 if (sum < ZERO_TOLERANCE) 67 { 68 curv= NAN_VALUE; 69 } 70 71 double eMin = std::min(std::min(e0,e1),e2); 72 curv= static_cast<ScalarType>(fabs(eMin) / sum); 73 74 75 for (unsigned i=0; i<n; ++i) 76 { 77 //current point index 78 unsigned index = cell.points->getPointGlobalIndex(i); 79 //current point index in neighbourhood (to compute curvature at the right position!) 80 unsigned indexInNeighbourhood = 0; 81 82 cell.points->setPointScalarValue(i,curv); 83 if (nProgress && !nProgress->oneStep()) 84 { 85 return false; 86 } 87 } 88 89 return true; 90 }
作者:太一吾鱼水
文章未经说明均属原创,学习笔记可能有大段的引用,一般会注明参考文献。
欢迎大家留言交流,转载请注明出处。
分类:
PointCloud
【推荐】国内首个AI IDE,深度理解中文开发场景,立即下载体验Trae
【推荐】编程新体验,更懂你的AI,立即体验豆包MarsCode编程助手
【推荐】抖音旗下AI助手豆包,你的智能百科全书,全免费不限次数
【推荐】轻量又高性能的 SSH 工具 IShell:AI 加持,快人一步
· 基于Microsoft.Extensions.AI核心库实现RAG应用
· Linux系列:如何用heaptrack跟踪.NET程序的非托管内存泄露
· 开发者必知的日志记录最佳实践
· SQL Server 2025 AI相关能力初探
· Linux系列:如何用 C#调用 C方法造成内存泄露
· 震惊!C++程序真的从main开始吗?99%的程序员都答错了
· 【硬核科普】Trae如何「偷看」你的代码?零基础破解AI编程运行原理
· 单元测试从入门到精通
· 上周热点回顾(3.3-3.9)
· winform 绘制太阳,地球,月球 运作规律