[{"data":1,"prerenderedAt":345},["ShallowReactive",2],{"method-han2026nifcyl":3},{"method":4,"reference":51,"equipment":71,"figures":97,"results":137},{"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":23,"limitations":29,"sensors":37,"platform":39,"estimator":41,"association":42,"timeModel":43,"deskew":43,"loopClosure":43,"globalOptimization":43,"mapRepresentation":44,"prior":45,"outputGeometry":46,"compute":47,"codeUrl":48,"codeLicense":49,"relatedVersions":50},"han2026nifcyl","Han et al., 2026","NIFCyl tunnel deformation from SLAM LiDAR","Full-field deformation quantification of underground tunnels using SLAM LiDAR point cloud based on unsupervised neural implicit learning",2026,"recent","C11b","downstream_engineering_task","作者提出 NIFCyl：以 8 層 MLP 非監督學習參考點雲的有號距離場，取其梯度作為尺度不變且方向一致的法向，再沿法向以圓柱鄰域平均兩期點雲的投影位置，求得全場變形，不需標註資料或局部 PCA 擬合。資料以手持 Hovermap ST 在西澳 Kalgoorlie 地下硬岩礦取得，兩期點雲以四個噴漆控制點做 SVD 粗配準，再以 ICP 細配準。合成變形情境中 NIFCyl 的 R² 為 0.963、RMSE 0.009 m、總計算 187 秒，三種尺度設定的 M3C2 中最佳者為 0.955、0.010 m、995 秒；現地放置紙箱的試驗中，平面區兩者 MAE 同為 0.022 m，曲面區為 0.032 m 對 0.034 m。實際擴挖案例只以單一點的人工量測（2.513 m）與模型結果（2.505 m）比對。","NIFCyl learns an SDF of the reference tunnel cloud with an MLP, uses its gradient as normals and averages cylinder projections along them; on handheld Hovermap ST scans from a Kalgoorlie mine it beats multi-scale M3C2 on synthetic deformation (R2 0.963, RMSE 0.009 m, 187 s vs 995 s) and on a curved field target (MAE 0.032 vs 0.034 m), and ties on a planar one (0.022 m).","full_text_reviewed","peer_reviewed_published","supplementary","西澳 Kalgoorlie 地下硬岩礦的鑽炸巷道（岩面粗糙，鋼網與岩栓支撐），含相隔兩個月、因撬落浮石（scaling）造成約 2 m 以上擴挖的實例；與隧道開挖變形監測高度相關，但屬礦業巷道而非土木隧道施工（Sec. 2.2, 4.5）。",[20,21,22],"simulation","controlled_experiment","underground_or_tunnel",[24,25,26,27,28],"synthetic: R2 0.963 and RMSE 0.009 m vs best M3C2 (0.05 to 0.55 m scale) 0.955 and 0.010 m; 187 s vs 995 s (Table 4)","curved field target MAE 0.032 m vs 0.034 m for M3C2 (Sec. 4.4)","robust to 20% point removal (R2 0.956, RMSE 0.010 m) and to noise at 10% of deformation (R2 0.938, RMSE 0.012 m) (Table 5)","noise floor in stable regions: sigma 0.0035 m short-term field, 0.0151 m over two months (Table 6)","code and datasets shared on GitHub (Data availability)",[30,31,32,33,34,35,36],"evaluation relies on a limited set of multi-temporal scans from one mine (Sec. 4.8)","ICP-based registration contributes systemic noise; real-scenario detection threshold set at 3.0 cm (Sec. 4.7, 4.8)","tunnel treated as a single entity without semantic segmentation (Sec. 4.8)","for tunnels longer than about 60 m the authors recommend splitting the cloud into sections (sliding window) because a single global volume causes memory and time bottlenecks (Sec. 4.3)","planar field target shows no gain over M3C2 (both MAE 0.022 m) (Sec. 4.4)","real-deformation case checked only qualitatively plus one manual point-to-point measurement (Sec. 4.5)","at noise 20% of deformation R2 drops to 0.858 and RMSE rises to 0.018 m (Table 5)",[38],"handheld Hovermap ST LiDAR SLAM scanner (Table 1: FoV 360 x 290 deg, range 0.40 to 100 m, LiDAR accuracy +\u002F-30 mm, mapping accuracy +\u002F-15 mm in typical underground and indoor environments, SLAM drift +\u002F-0.03%, up to 300,000 pts\u002Fs single return and 600,000 pts\u002Fs dual return)",[40],"handheld","not_applicable (deformation analysis of SLAM point clouds)","SDF learned by an 8-layer MLP gives normals via its gradient; deformation measured by averaging reference and compared points inside a cylinder along each normal; epochs aligned by GCP-based SVD then ICP","not_applicable","neural implicit function of tunnel surface; scale-invariant normal field (per abstract)","no labels (unsupervised); four painted ground control points per scan location used for coarse registration of epochs before ICP","full-field deformation map","Linux desktop with Intel Core i9 14900K CPU and NVIDIA RTX4080 Super GPU; NIFCyl 187 s total (47 s deformation step) vs M3C2 995 to 6276 s on the synthetic set; 20,000 iterations, Adam",null,"not_verified",[],{"id":5,"kind":52,"shortName":7,"title":8,"authors":53,"year":9,"venue":58,"venueType":59,"publisher":60,"volumeIssuePages":61,"doi":62,"arxivId":48,"url":63,"firstPublicDate":64,"publicationStatus":16,"metadataStatus":65,"fulltextStatus":15,"era":10,"classicReason":43,"codeUrl":48,"cluster":11,"topics":66,"mdpi":67,"verification":68,"label":6,"fulltextRoute":69,"versionRead":70,"addedByCensus":67},"method",[54,55,56,57],"Zhen Han","Wei Lin","Qian Li","Danqi Li","Tunnelling and Underground Space Technology","journal","Elsevier","175, 107777","10.1016\u002Fj.tust.2026.107777","https:\u002F\u002Fapi.openalex.org\u002Fworks\u002Fdoi:10.1016\u002Fj.tust.2026.107777","2026-05-11","metadata_verified",[11],false,"confirmed","publisher OA","version of record, Tunnelling and Underground Space Technology 175:107777 (September 2026), ScienceDirect HTML full text, open access under CC BY 4.0",[72,78,84,87,92],{"category":73,"model":74,"canonical":74,"role":75,"dataset":48,"specs":76,"locator":77},"mobile_scanner_device","Hovermap ST (Table 1: 'Hovermap ST AUTONOMY')","method input","SLAM drift +\u002F-0.03%; FoV 360 x 290 deg; range 0.40 to 100 m; LiDAR accuracy +\u002F-30 mm; mapping accuracy +\u002F-15 mm in typical underground and indoor environments; intensity, range, time, return number and ring attributes; up to 300,000 pts\u002Fs single return, 600,000 pts\u002Fs dual return; handheld, 60 m closed loop in about 5 min","Sec. 2.2, Table 1",{"category":79,"model":80,"canonical":80,"role":81,"dataset":48,"specs":82,"locator":83},"compute","Intel Core i9 14900K","compute for runtime","desktop CPU, Linux","Sec. 4.1",{"category":79,"model":85,"canonical":85,"role":81,"dataset":48,"specs":86,"locator":83},"NVIDIA RTX4080 Super","desktop GPU",{"category":88,"model":89,"canonical":89,"role":75,"dataset":48,"specs":90,"locator":91},"other","painted ground control points (circular marker with crosshair on white spray coating, Roman numeral ID)","four stable sidewall regions per scan location","Sec. 2.1, Fig. 1",{"category":88,"model":93,"canonical":93,"role":94,"dataset":48,"specs":95,"locator":96},"ruler","reference or ground truth","measured box height 9.5 cm as planar ground truth","Sec. 4.4",[98,111,121,129],{"refId":5,"refLabel":6,"fig":99,"whatZh":100,"license":101,"licenseUrl":102,"sourceUrl":103,"src":104,"width":105,"height":106,"thumb":107,"thumbWidth":108,"thumbHeight":109,"modified":110},"Fig. 2","地下硬岩礦巷道現場環境：低光、多粉塵、鋼網與岩栓支撐的粗糙岩面","CC BY 4.0","https:\u002F\u002Fcreativecommons.org\u002Flicenses\u002Fby\u002F4.0\u002F","https:\u002F\u002Fars.els-cdn.com\u002Fcontent\u002Fimage\u002F1-s2.0-S0886779826003354-gr2_lrg.jpg","\u002Ffigure-files\u002Fhan2026nifcyl\u002Ffig-2.webp",1367,804,"\u002Ffigure-files\u002Fhan2026nifcyl\u002Ffig-2.thumb.webp",480,282,"converted to WebP",{"refId":5,"refLabel":6,"fig":112,"whatZh":113,"license":101,"licenseUrl":102,"sourceUrl":114,"src":115,"width":116,"height":117,"thumb":118,"thumbWidth":108,"thumbHeight":119,"modified":120},"Fig. 3","NIFCyl 方法流程：合成變形資料、神經隱式法向計算、圓柱式變形量測","https:\u002F\u002Fars.els-cdn.com\u002Fcontent\u002Fimage\u002F1-s2.0-S0886779826003354-gr3_lrg.jpg","\u002Ffigure-files\u002Fhan2026nifcyl\u002Ffig-3.webp",1400,740,"\u002Ffigure-files\u002Fhan2026nifcyl\u002Ffig-3.thumb.webp",254,"resized to at most 1400 px wide and converted to WebP",{"refId":5,"refLabel":6,"fig":122,"whatZh":123,"license":101,"licenseUrl":102,"sourceUrl":124,"src":125,"width":116,"height":126,"thumb":127,"thumbWidth":108,"thumbHeight":128,"modified":120},"Fig. 4","巷道點雲（約 185 萬點）中以數學曲面取代的五個測試區 A 至 E，及單一區域俯視與側視","https:\u002F\u002Fars.els-cdn.com\u002Fcontent\u002Fimage\u002F1-s2.0-S0886779826003354-gr4_lrg.jpg","\u002Ffigure-files\u002Fhan2026nifcyl\u002Ffig-4.webp",1257,"\u002Ffigure-files\u002Fhan2026nifcyl\u002Ffig-4.thumb.webp",431,{"refId":5,"refLabel":6,"fig":130,"whatZh":131,"license":101,"licenseUrl":102,"sourceUrl":132,"src":133,"width":116,"height":134,"thumb":135,"thumbWidth":108,"thumbHeight":136,"modified":120},"Fig. 15","實際擴挖案例的 NIFCyl 結果：大於 2 m 擴挖的點以紅色標示，並以兩個平坦區中心距離作人工比對","https:\u002F\u002Fars.els-cdn.com\u002Fcontent\u002Fimage\u002F1-s2.0-S0886779826003354-gr15_lrg.jpg","\u002Ffigure-files\u002Fhan2026nifcyl\u002Ffig-15.webp",625,"\u002Ffigure-files\u002Fhan2026nifcyl\u002Ffig-15.thumb.webp",214,{"totalRows":138,"groupCount":139,"groups":140,"others":339},23,5,[141,208,253,309],{"slug":142,"group":143,"sourceId":5,"sourceLabel":6,"table":144,"selfRows":145,"metrics":146,"seqs":162,"entrants":175,"cells":179,"outcomes":201,"locators":203,"hardware":204,"wordings":205,"notes":206},"han2026nifcyl-table-6","han2026nifcyl:Table 6","Table 6",9,[147,152,155,158,160,161],{"label":148,"unit":149,"statistic":150,"alignment":151},"u (m), Gaussian fitted mean residual","m","mean","none",{"label":153,"unit":149,"statistic":154,"alignment":151},"sigma (m), Gaussian fitted standard deviation","std",{"label":156,"unit":149,"statistic":157,"alignment":151},"95% confidence (1.96 sigma) detection threshold (m)","not_reported",{"label":148,"unit":149,"statistic":150,"alignment":159},"control points",{"label":153,"unit":149,"statistic":154,"alignment":159},{"label":156,"unit":149,"statistic":157,"alignment":159},[163,167,171],{"dataset":164,"sequence":165,"environment":166},"synthetic deformation benchmark","synthetic deformation, stable region","underground mine drive (synthetic)",{"dataset":168,"sequence":169,"environment":170},"own field test, Kalgoorlie underground mine","experimental deformation scans (30 min apart), stable region","underground hard rock mine drive",{"dataset":172,"sequence":173,"environment":174},"own field data, Kalgoorlie underground mine","real deformation scans (two months apart), floor region","active underground mine drive",[176],{"name":177,"methodId":5,"linkable":178,"proposed":178,"self":178},"NIFCyl",true,[180,183,185,187,190,193,195,197,199],[181,181,181,181,182,181,182,182,181],0,-1,[181,184,181,181,182,181,182,182,181],1,[181,186,181,181,182,181,182,182,181],2,[181,188,184,189,182,181,182,182,181],3,0.0007,[181,191,184,192,182,181,182,182,181],4,0.0035,[181,139,184,194,182,181,182,182,181],0.0069,[181,188,186,196,181,181,182,182,181],-0.0007,[181,191,186,198,182,181,182,182,181],0.0151,[181,139,186,200,182,181,182,182,181],0.0296,[202],"text in Sec. 4.7 gives +0.0007 m",[144],[],[],[207],"Deformation residuals of NIFCyl in undeformed regions (Gaussian fit); synthetic N=113,059, experimental N=121,151, real N=104,031",{"slug":209,"group":210,"sourceId":5,"sourceLabel":6,"table":211,"selfRows":212,"metrics":213,"seqs":220,"entrants":229,"cells":231,"outcomes":247,"locators":248,"hardware":249,"wordings":250,"notes":251},"han2026nifcyl-table-5","han2026nifcyl:Table 5","Table 5",8,[214,217],{"label":215,"unit":216,"statistic":157,"alignment":151},"R2","unitless",{"label":218,"unit":149,"statistic":219,"alignment":151},"RMSE (m)","RMSE",[221,223,225,227],{"dataset":164,"sequence":222,"environment":166},"point density -10%",{"dataset":164,"sequence":224,"environment":166},"point density -20%",{"dataset":164,"sequence":226,"environment":166},"Gaussian noise +10% sigma",{"dataset":164,"sequence":228,"environment":166},"Gaussian noise +20% sigma",[230],{"name":177,"methodId":5,"linkable":178,"proposed":178,"self":178},[232,234,236,238,239,241,243,245],[181,181,181,233,182,181,182,182,181],0.959,[181,184,181,235,182,181,182,182,181],0.01,[181,181,184,237,182,181,182,182,181],0.956,[181,184,184,235,182,181,182,182,181],[181,181,186,240,182,181,182,182,181],0.938,[181,184,186,242,182,181,182,182,181],0.012,[181,181,188,244,182,181,182,182,181],0.858,[181,184,188,246,182,181,182,182,181],0.018,[],[211],[],[],[252],"Robustness of NIFCyl on the synthetic dataset under random point removal and added zero-mean Gaussian noise (sigma as % of deformation)",{"slug":254,"group":255,"sourceId":5,"sourceLabel":6,"table":256,"selfRows":188,"metrics":257,"seqs":265,"entrants":270,"cells":279,"outcomes":302,"locators":303,"hardware":304,"wordings":306,"notes":307},"han2026nifcyl-table-4","han2026nifcyl:Table 4","Table 4",[258,260,262],{"label":259,"unit":216,"statistic":157,"alignment":151},"R2 (against the 1:1 line)",{"label":261,"unit":149,"statistic":219,"alignment":151},"RMSE (m) of deformation vs ground truth",{"label":263,"unit":264,"statistic":157,"alignment":151},"Computing Time (s)","s",[266],{"dataset":267,"sequence":268,"environment":269},"synthetic deformation benchmark from Hovermap ST scan","regions A to E","underground mine drive (5 x 4.5 x 8 m)",[271,272,275,277],{"name":177,"methodId":5,"linkable":178,"proposed":178,"self":178},{"name":273,"methodId":274,"linkable":67,"proposed":67,"self":67},"M3C2 small scale (0.05 to 0.55 m)","lague2013m3c2",{"name":276,"methodId":274,"linkable":67,"proposed":67,"self":67},"M3C2 medium scale (0.55 to 1.05 m)",{"name":278,"methodId":274,"linkable":67,"proposed":67,"self":67},"M3C2 large scale (1.05 to 1.55 m)",[280,282,284,286,288,289,291,293,294,296,298,300],[181,181,181,281,182,181,182,182,181],0.963,[181,184,181,283,182,181,182,182,181],0.009,[181,186,181,285,182,181,181,182,181],187,[184,181,181,287,182,181,182,182,181],0.955,[184,184,181,235,182,181,182,182,181],[184,186,181,290,182,181,181,182,181],995,[186,181,181,292,182,181,182,182,181],0.937,[186,184,181,242,182,181,182,182,181],[186,186,181,295,182,181,181,182,181],3601,[188,181,181,297,182,181,182,182,181],0.931,[188,184,181,299,182,181,182,182,181],0.013,[188,186,181,301,182,181,181,182,181],6276,[],[256],[305],"Linux desktop, Intel Core i9 14900K CPU, NVIDIA RTX4080 Super GPU (Sec. 4.1, stated for all experiments)",[],[308],"Synthetic deformation on 5 regions (88,842 points) of a real mine tunnel cloud; M3C2 at three normal-scale ranges with empty values excluded",{"slug":310,"group":311,"sourceId":5,"sourceLabel":6,"table":312,"selfRows":186,"metrics":313,"seqs":316,"entrants":321,"cells":325,"outcomes":333,"locators":334,"hardware":335,"wordings":336,"notes":337},"han2026nifcyl-text-sec-4-4","han2026nifcyl:Text Sec.4.4","Text Sec.4.4",[314],{"label":315,"unit":149,"statistic":150,"alignment":157},"Mean Absolute Error (m)",[317,319],{"dataset":168,"sequence":318,"environment":170},"plane surface deformation region",{"dataset":168,"sequence":320,"environment":170},"curved (spherical) surface deformation region",[322,323],{"name":177,"methodId":5,"linkable":178,"proposed":178,"self":178},{"name":324,"methodId":274,"linkable":67,"proposed":67,"self":67},"M3C2",[326,328,329,331],[181,181,181,327,182,181,182,182,181],0.022,[184,181,181,327,182,181,182,182,181],[181,181,184,330,182,181,182,182,181],0.032,[184,181,184,332,182,181,182,182,181],0.034,[],[96],[],[],[338],"Field test with two scans within half an hour; paper box (9.5 cm) on flat cardboard and box over half basketball on cardboard; Sec. 4.4 does not state how the two scans were aligned (the general Sec. 2.3 pipeline is GCP-based SVD then ICP)",[340],{"group":341,"slug":342,"sourceLabel":6,"table":343,"selfRows":184,"datasets":344},"han2026nifcyl:Text Sec.4.5","han2026nifcyl-text-sec-4-5","Text Sec.4.5",[172],1790510659181]