[{"data":1,"prerenderedAt":128},["ShallowReactive",2],{"method-rusu2011pcl":3},{"method":4,"reference":39,"equipment":59,"figures":72,"results":73},{"id":5,"label":6,"shortName":7,"title":8,"year":9,"era":10,"cluster":11,"scope":12,"keyIdeaZh":13,"keyIdeaEn":14,"fulltextStatus":15,"publicationStatus":16,"recommendation":17,"constructionRelevance":18,"validationEnvironment":19,"strengths":20,"limitations":25,"sensors":29,"platform":30,"estimator":31,"association":31,"timeModel":31,"deskew":31,"loopClosure":31,"globalOptimization":32,"mapRepresentation":33,"prior":32,"outputGeometry":34,"compute":35,"codeUrl":36,"codeLicense":37,"relatedVersions":38},"rusu2011pcl","Rusu & Cousins, 2011","PCL","3D is here: Point Cloud Library (PCL)",2011,"classic","C12","sensing_calibration_sync_preprocessing","PCL 是以 C++ 模板實作、採 BSD 授權的開源點雲處理函式庫，分成濾波、特徵、輸入輸出、分割、表面重建、配準、關鍵點與距離影像等可獨立編譯的模組，底層以 Eigen、FLANN 與 OpenMP 或 TBB 支援線性代數、近鄰搜尋與多核心平行化，並以 ROS nodelet 在同一行程內串接處理圖，避免資料複製。文中示範統計離群值移除（k = 50、1.0σ）與 RANSAC 平面分割（1 cm 門檻）。PCL 的體素格網濾波常用於 SLAM 前處理降採樣，此點屬推論，本文沒有討論，也未評估降採樣對幾何的影響。","An open-source, BSD-licensed C++ library for point-cloud processing covering filtering, features, reconstruction, registration and segmentation.","full_text_reviewed","peer_reviewed_published","supplementary","本論文未在營建場域驗證，也沒有量化評估；文中以 PR2 機器人的門與把手辨識等室內示例說明用途，並指出牆面、門與桌面偵測可共用同一個受限平面分割模組，只需調整參數（Sec. III），這與營建構件的平面擷取相關（推論）。「被大量用於前處理」屬外部觀察，本文沒有提供證據。",[],[21,22,23,24],"Modular pipeline interface and multi-core support (Sec. II).","Unit and regression tests compiled and checked on a dedicated build farm (Sec. II).","ROS nodelets connect processing blocks inside one process without copying or serialising point clouds (Sec. III).","Ported to Windows, MacOS and Linux as of version 0.6 (Sec. II).",[26,27,28],"Open3D authors describe PCL as encumbered by bloat and largely dormant as of 2018 (Zhou et al. 2018, Sec. 1); this is a competing-library claim, not an independent evaluation.","At the time of writing (version 0.6, before 1.0) GPU support through CUDA or OpenCL was only planned, and documentation, unit tests and tutorials were still to be improved (Sec. VI).","No quantitative accuracy or runtime evaluation; the usage examples are qualitative (Sec. V).",[],[],"not_applicable","none","point clouds (templated n-D point types)","library operations incl. filtering (e.g., voxel grid), features, registration, surface reconstruction","CPU with SSE, OpenMP\u002FTBB parallelization","https:\u002F\u002Fgithub.com\u002FPointCloudLibrary\u002Fpcl","BSD (repository LICENSE.txt)",[],{"id":5,"kind":40,"shortName":7,"title":8,"authors":41,"year":9,"venue":44,"venueType":45,"publisher":46,"volumeIssuePages":47,"doi":48,"arxivId":49,"url":50,"firstPublicDate":51,"publicationStatus":16,"metadataStatus":52,"fulltextStatus":15,"era":10,"classicReason":53,"codeUrl":36,"cluster":11,"topics":54,"mdpi":55,"verification":56,"label":6,"fulltextRoute":57,"versionRead":58,"addedByCensus":55},"software",[42,43],"Radu Bogdan Rusu","Steve Cousins","2011 IEEE International Conference on Robotics and Automation (ICRA)","conference","IEEE","pp. 1-4","10.1109\u002Ficra.2011.5980567",null,"https:\u002F\u002Fpointclouds.org\u002Fassets\u002Fpdf\u002Fpcl_icra2011.pdf","2011-05","metadata_verified","reproducible baseline: widely used implementations of voxel-grid filtering, kd-tree search and surface reconstruction; ikd-Tree benchmarks against PCL's static kd-tree and ImMesh uses PCL's TSDF as a baseline.",[11],false,"corrected","author copy","Author-hosted camera-ready PDF on pointclouds.org (4 pp., pdfTeX, created 2011-03-20) read in full; the IEEE Xplore version of record (pp. 1-4) was not compared in this pass",[60,66],{"category":61,"model":62,"canonical":62,"role":63,"dataset":49,"specs":64,"locator":65},"lidar","tilting laser scanner (model not named)","dataset sensor","source of the raw point cloud in the statistical outlier removal example (Fig. 7); the paper does not say which robot carried it","Fig. 7 caption",{"category":67,"model":68,"canonical":68,"role":69,"dataset":49,"specs":70,"locator":71},"platform","PR2 robot","method input","platform of the door and handle identification, NARF object recognition and grasping examples, which are reproduced from cited works [8]-[11]","Sec. V; Figs. 9-10",[],{"totalRows":74,"groupCount":75,"groups":76,"others":127},4,1,[77],{"slug":78,"group":79,"sourceId":80,"sourceLabel":81,"table":82,"selfRows":74,"metrics":83,"seqs":88,"entrants":96,"cells":104,"outcomes":120,"locators":121,"hardware":122,"wordings":124,"notes":125},"loglio2024-table-i","loglio2024:Table I","loglio2024","Huang et al., 2024b","Table I",[84],{"label":85,"unit":86,"statistic":87,"alignment":31},"mean running time of normal estimation","ms","mean",[89,93],{"dataset":90,"sequence":91,"environment":92},"M2DGR","Velodyne-32 (57,600 points)","per LiDAR scan",{"dataset":94,"sequence":95,"environment":92},"NTU VIRAL","Ouster-16 (16,384 points)",[97,100,102],{"name":98,"methodId":80,"linkable":99,"proposed":99,"self":55},"Ring FALS (total)",true,{"name":101,"methodId":5,"linkable":99,"proposed":55,"self":99},"PCL single thread",{"name":103,"methodId":5,"linkable":99,"proposed":55,"self":99},"PCL OMP 10 threads",[105,109,111,114,116,118],[106,106,106,107,108,106,106,108,106],0,7.784,-1,[75,106,106,110,108,106,106,108,106],79.811,[112,106,106,113,108,106,106,108,106],2,26.355,[106,106,75,115,108,106,106,108,106],2.597,[75,106,75,117,108,106,106,108,106],155.972,[112,106,75,119,108,106,106,108,106],39.664,[],[82],[123],"Intel Xeon Gold 6248R 3.00 GHz, 32 GB RAM",[],[126],"Mean running time of normal estimation for a single scan; Ring FALS includes projection, box-filtering and smoothing; PCL least squares with k-d tree, without smoothing",[],1790510665529]