[{"data":1,"prerenderedAt":276},["ShallowReactive",2],{"method-scancontext2018":3},{"method":4,"reference":59,"equipment":78,"figures":106,"results":107},{"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":22,"limitations":27,"sensors":35,"platform":37,"estimator":41,"association":42,"timeModel":41,"deskew":43,"loopClosure":44,"globalOptimization":45,"mapRepresentation":46,"prior":47,"outputGeometry":41,"compute":48,"codeUrl":49,"codeLicense":50,"relatedVersions":51},"scancontext2018","Kim & Kim, 2018","Scan Context","Scan Context: Egocentric Spatial Descriptor for Place Recognition Within 3D Point Cloud Map",2018,"recent","C06","place_recognition_component","Scan Context 以感測器為中心，將單次 3D 光達掃描劃分為 20 個環（ring）乘 60 個扇區（sector）的極座標格網（最大距離 80 m），每格記錄其中點的最大高度，形成 2D 全域描述子，不依賴直方圖或事前訓練。搜尋分兩階段：先以各環佔有率組成的環鍵（ring key）建 kd-tree 取出 10 或 50 個候選，再對候選做所有欄位平移的逐欄餘弦距離比較，最小距離低於門檻即判定為迴圈；平移量同時給出約 6° 解析度的偏航初值，可供 ICP 使用，因此反向重訪與轉角也能偵測迴圈。","Scan Context encodes a LiDAR scan as an egocentric polar grid descriptor and detects loops by ring-key kd-tree retrieval plus column-shift (yaw-aligned) comparison, without training.","full_text_reviewed","peer_reviewed_published","main_body","未報告工地測試。原論文指出在含狹窄室內空間的 NCLT 路段，因垂直高度變化小而召回率與精確率偏低（Sec. IV-B），此點與施工中建築的走廊及樓層環境相關，但套用到工地屬推論。LT-mapper 以 Scan Context 作跨時段迴圈（ltmapper2022 Sec. V-A1），BIM-SLAM 亦採用並調整室內參數（bimslam2023 Sec. 4.3），因此它是營建相關多時段流程的實際依賴元件。",[20,21],"public_benchmark","cross_site",[23,24,25,26],"Viewpoint (reverse revisit) invariance through column-shift alignment; substantially outperformed M2DP, Z-projection and ESF on KITTI 08, which has only reverse loops (Sec. IV-B, Fig. 5d)","No histogram and no training required (abstract)","Authors report outperforming M2DP, Z-projection and ESF on the outdoor urban sequences, including the challenging Complex Urban 02, in precision-recall curves (Sec. IV-B, Fig. 5)","Yaw estimate from column shift improves ICP success, time and RMSE on reverse loops (Sec. IV-C, Fig. 7, 8)",[28,29,30,31,32,33,34],"Authors report limited performance indoors where vertical height variation is small: low recall and precision on NCLT, whose trajectory contains narrow indoor spaces (Sec. IV-B)","All methods degraded on Complex Urban 02 with narrow roads and repeated structures of similar height and shape (Sec. IV-B, Fig. 6)","Ring key is less informative than the full descriptor, so only 10 candidates is vulnerable with many similar structures (Sec. IV-B)","Loop search slower than all three compared global descriptors (M2DP, Z-projection, ESF), though within 2 to 5 Hz in Matlab (Table II, Sec. IV-D)","Code is CC BY-NC-SA 4.0, i.e., non-commercial use only (README)","Found in the authors' journal extension (vehicle driving data): a vehicle moving along a corridor-like place and a tall, large object (e.g., a bus) very close to the sensor caused localization failures (scancontextpp2022 Sec. VIII-F, Fig. 22)","Found in follow-up work: LT-SLAM with Scan Context failed to find loops in a repetitive indoor parking dataset (yang2024lifelong Sec. IV-B)",[36],"3D LiDAR (Velodyne HDL-64E on KITTI, HDL-32E on NCLT, two tilted VLP-16 merged on Complex Urban LiDAR)",[38,39,40],"vehicle (KITTI: HDL-64E located in the centre of the car)","Segway mobile platform (NCLT)","Complex Urban LiDAR dataset (two tilted VLP-16; platform not described in the paper)","not_applicable","egocentric polar grid Nr = 20 rings x Ns = 60 sectors, Lmax = 80 m, bin value = maximum point height (Eq. 3), empty bins 0; optional root-shift augmentation with Ntrans = 8 translated copies for lane-level offsets; ring key = per-ring occupancy ratio (L0 norm) indexed in a KD tree, 10 or 50 candidates; column-wise cosine distance minimised over all column shifts, accepted below threshold tau; 0.6 m grid downsampling","not_reported","provides loop candidates with coarse yaw alignment (6 deg resolution); the yaw shift initialises point-to-point ICP, which reduced ICP time and RMSE for KITTI 08 reverse loops (Fig. 7, 8); pose-graph use left to host SLAM","none (component)","per-keyframe 2D descriptor database","none; no training required (abstract)","Matlab on Intel i7-6700 3.40 GHz with 16 GB memory; on KITTI 00, 0.1291 s per descriptor with augmentation (0.0143 s without) and 0.0807 s or 0.3331 s loop search for 10 or 50 candidates; about 2 to 5 Hz overall (Table II, Sec. IV-D)","https:\u002F\u002Fgithub.com\u002Fgisbi-kim\u002Fscancontext","CC BY-NC-SA 4.0 (stated in README; non-commercial)",[52,56],{"relation":53,"title":54,"doi_or_url":55},"journal_extension","Scan Context++: Structural Place Recognition Robust to Rotation and Lateral Variations in Urban Environments (T-RO 2022)","10.1109\u002FTRO.2021.3116424",{"relation":57,"title":58,"doi_or_url":49},"code_release","gisbi-kim\u002Fscancontext",{"id":5,"kind":60,"shortName":7,"title":8,"authors":61,"year":9,"venue":64,"venueType":65,"publisher":66,"volumeIssuePages":67,"doi":68,"arxivId":69,"url":70,"firstPublicDate":71,"publicationStatus":16,"metadataStatus":72,"fulltextStatus":15,"era":10,"classicReason":41,"codeUrl":49,"cluster":11,"topics":73,"mdpi":74,"verification":75,"label":6,"fulltextRoute":76,"versionRead":77,"addedByCensus":74},"component",[62,63],"Giseop Kim","Ayoung Kim","2018 IEEE\u002FRSJ International Conference on Intelligent Robots and Systems (IROS)","conference","IEEE","pp. 4802-4809","10.1109\u002Firos.2018.8593953",null,"https:\u002F\u002Fdoi.org\u002F10.1109\u002FIROS.2018.8593953","2018-10","metadata_verified",[11],false,"corrected","author copy","Author copy PDF (gisbi-kim.github.io\u002Fpublications\u002Fgkim-2018-iros.pdf, 8 pages, accepted-manuscript layout) read in full; cross-checked against the IEEE Xplore version of record HTML (document 8593953): same section structure, Table II image identical, key text (0.0143 s, i7-6700, root shifting, indoor limitation, 2-5 Hz) present",[79,86,91,97,100],{"category":80,"model":81,"canonical":81,"role":82,"dataset":83,"specs":84,"locator":85},"lidar","Velodyne HDL-64E","dataset sensor","KITTI","64-ray, located in the centre of the car","Sec. IV-A1",{"category":80,"model":87,"canonical":87,"role":82,"dataset":88,"specs":89,"locator":90},"Velodyne HDL-32E","NCLT","32-ray, attached to a Segway mobile platform","Sec. IV-A2",{"category":80,"model":92,"canonical":93,"role":82,"dataset":94,"specs":95,"locator":96},"Velodyne VLP-16 PUCK (two, tilted left and right)","Velodyne VLP-16","Complex Urban LiDAR","two tilted units without 360 deg surround view; clouds merged into one scan","Sec. IV-A3",{"category":98,"model":99,"canonical":99,"role":82,"dataset":88,"specs":43,"locator":90},"platform","Segway mobile platform",{"category":101,"model":102,"canonical":102,"role":103,"dataset":69,"specs":104,"locator":105},"compute","Intel i7-6700 CPU","compute for runtime","3.40 GHz, 16 GB memory; Matlab","Sec. IV",[],{"totalRows":108,"groupCount":109,"groups":110,"others":275},11,3,[111,199,253],{"slug":112,"group":113,"sourceId":114,"sourceLabel":115,"table":116,"selfRows":117,"metrics":118,"seqs":124,"entrants":133,"cells":145,"outcomes":193,"locators":194,"hardware":195,"wordings":196,"notes":197},"iscloam2020-table-ii","iscloam2020:Table II","iscloam2020","Wang et al., 2020","Table II",6,[119,122],{"label":120,"unit":121,"statistic":43,"alignment":41},"Precision (%)","%",{"label":123,"unit":121,"statistic":43,"alignment":41},"Recall Rate (%)",[125,129,131],{"dataset":126,"sequence":127,"environment":128},"KITTI odometry","sequence 00","vehicle, urban and residential",{"dataset":126,"sequence":130,"environment":128},"sequence 02",{"dataset":126,"sequence":132,"environment":128},"sequence 05",[134,137,139,141,143],{"name":135,"methodId":5,"linkable":136,"proposed":74,"self":136},"Kim [21] (Scan Context)",true,{"name":138,"methodId":69,"linkable":74,"proposed":74,"self":74},"GLAROT3D [17]",{"name":140,"methodId":69,"linkable":74,"proposed":74,"self":74},"Cieslewski [24]",{"name":142,"methodId":69,"linkable":74,"proposed":74,"self":74},"Galvez-Lopez [10] (DBoW2)",{"name":144,"methodId":114,"linkable":136,"proposed":136,"self":74},"Proposed (ISC)",[146,150,153,155,157,160,162,163,164,166,168,170,172,173,175,177,179,180,181,183,185,186,187,188,190,191],[147,147,147,148,149,147,149,149,147],0,100,-1,[147,151,147,152,149,147,149,149,147],1,87,[151,147,147,154,149,147,149,149,147],86,[151,151,147,156,149,147,149,149,147],40,[158,147,147,159,149,147,149,149,147],2,92,[158,151,147,161,149,147,149,149,147],80,[109,147,147,148,149,147,149,149,147],[109,151,147,159,149,147,149,149,147],[165,147,147,148,149,147,149,149,147],4,[165,151,147,167,149,147,149,149,147],90.2,[147,147,151,169,149,147,149,149,147],90,[147,151,151,171,149,147,149,149,147],73,[109,147,151,148,149,147,149,149,147],[109,151,151,174,149,147,149,149,147],80.6,[165,147,151,176,149,147,149,149,147],98,[165,151,151,178,149,147,149,149,147],91,[147,147,158,148,149,147,149,149,147],[147,151,158,169,149,147,149,149,147],[158,147,158,182,149,147,149,149,147],93,[158,151,158,184,149,147,149,149,147],60,[151,147,158,161,149,147,149,149,147],[151,151,158,161,149,147,149,149,147],[109,147,158,148,149,147,149,149,147],[109,151,158,189,149,147,149,149,147],87.6,[165,147,158,148,149,147,149,149,147],[165,151,158,192,149,147,149,149,147],91.2,[],[116],[],[],[198],"KITTI sequences 00, 02 (forward and reverse revisits), 05; loop-closure precision and recall (%); Scan Context, GLAROT3D and Cieslewski results copied from their papers, DBoW2 run by the authors; loop ground truth from GPS",{"slug":200,"group":201,"sourceId":5,"sourceLabel":6,"table":116,"selfRows":165,"metrics":202,"seqs":209,"entrants":213,"cells":225,"outcomes":245,"locators":246,"hardware":247,"wordings":250,"notes":251},"scancontext2018-table-ii","scancontext2018:Table II",[203,207],{"label":204,"unit":205,"statistic":206,"alignment":41},"Calculating Descriptor (s)","s","mean",{"label":208,"unit":205,"statistic":206,"alignment":41},"Searching Loop (s), includes KD tree creation and distance computation",[210],{"dataset":83,"sequence":211,"environment":212},"00","urban driving",[214,216,218,221,223],{"name":215,"methodId":5,"linkable":136,"proposed":136,"self":136},"Scan context-10",{"name":217,"methodId":5,"linkable":136,"proposed":136,"self":136},"Scan context-50",{"name":219,"methodId":220,"linkable":136,"proposed":74,"self":74},"M2DP","m2dp2016",{"name":222,"methodId":69,"linkable":74,"proposed":74,"self":74},"Z-projection",{"name":224,"methodId":69,"linkable":74,"proposed":74,"self":74},"ESF",[226,228,230,231,233,235,237,239,241,243],[147,147,147,227,149,147,147,149,147],0.1291,[147,151,147,229,149,147,147,149,147],0.0807,[151,147,147,227,149,147,147,149,147],[151,151,147,232,149,147,147,149,147],0.3331,[158,147,147,234,149,147,147,149,147],0.0218,[158,151,147,236,149,147,147,149,147],0.0032,[109,147,147,238,149,147,147,149,147],0.0472,[109,151,147,240,149,147,147,149,147],0.0035,[165,147,147,242,149,147,151,149,147],0.0635,[165,151,147,244,149,147,151,149,147],0.0043,[],[116],[248,249],"Intel i7-6700 CPU 3.40 GHz, 16 GB memory (Matlab)","Intel i7-6700 CPU 3.40 GHz, 16 GB memory (PCL C++)",[],[252],"Average time on KITTI 00; 0.6 m3 grid downsampling for all methods (Sec. IV-D); scan context creation includes optional root-shift augmentation; implementations: Scan Context in Matlab, M2DP authors' Matlab code, Z-projection in Matlab, ESF from PCL in C++",{"slug":254,"group":255,"sourceId":5,"sourceLabel":6,"table":256,"selfRows":151,"metrics":257,"seqs":260,"entrants":262,"cells":265,"outcomes":268,"locators":269,"hardware":271,"wordings":272,"notes":273},"scancontext2018-text-sec-iv-d","scancontext2018:Text Sec.IV-D","Text Sec.IV-D",[258],{"label":259,"unit":205,"statistic":206,"alignment":41},"time to create a single scan context (without augmentation)",[261],{"dataset":83,"sequence":211,"environment":212},[263],{"name":264,"methodId":5,"linkable":136,"proposed":136,"self":136},"Scan context",[266],[147,147,147,267,149,147,147,149,147],0.0143,[],[270],"Sec. IV-D",[248],[],[274],"Single scan context creation without augmentation",[],1790510654779]