[{"data":1,"prerenderedAt":306},["ShallowReactive",2],{"method-vloam2015":3},{"method":4,"reference":54,"equipment":73,"figures":111,"results":112},{"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":33,"platform":38,"estimator":41,"association":42,"timeModel":43,"deskew":44,"loopClosure":45,"globalOptimization":46,"mapRepresentation":47,"prior":48,"outputGeometry":49,"compute":50,"codeUrl":51,"codeLicense":52,"relatedVersions":53},"vloam2015","Zhang & Singh, 2015","V-LOAM","Visual-lidar odometry and mapping: low-drift, robust, and fast",2015,"classic","C07","odometry_with_local_mapping","V-LOAM 以單眼相機搭配掃描式 3D 光達（由馬達帶動的 Hokuyo 2D 雷射掃描儀），分成兩個依序運作的階段：視覺里程計以影像速率（60 Hz）估計相鄰影格間的運動，特徵點的深度取自光達深度圖或三角化，沒有深度的特徵也納入求解；光達里程計每次掃描（約 1 秒）執行一次，先以線性運動模型做掃描對掃描精修，消除視覺漂移造成的點雲畸變，再以邊緣與平面特徵做掃描對地圖配準並累積地圖。最後整合低頻光達位姿與高頻視覺運動，以影像速率輸出位姿。方法與硬體都沒有使用 IMU，也刻意不做迴圈閉合。","Visual odometry provides high-rate, low-fidelity motion for registering scanning-LiDAR points, which LiDAR scan-matching odometry then refines.","full_text_reviewed","peer_reviewed_published","background","未在施工工地測試。自建資料為手持感測器在建物室內、穿越建物的室內外路徑（538 m）、含七個 180 度轉彎的樓梯間與走廊的測試，另有關燈造成光照劇變的室內測試（Sec. VII）；並以 KITTI 車載資料驗證。精度以迴圈缺口、人工對應點或起終點差估算，沒有獨立參考量測。與營建的關聯僅在於樓梯間與走廊等既有建物情境。",[20,21],"public_benchmark","completed_building",[23,24,25,26],"Ranked first on the KITTI odometry benchmark at the time, 0.75% relative position drift (abstract, Sec. VIII)","Relative position error 0.31% to 0.73% on own handheld tests of 47 m to 538 m; wide-angle and fisheye setups reach the same level after lidar refinement (Table I)","With a fisheye camera, fast trials finished at 1.3% (staircase, up to about 170 deg\u002Fs) and 0.39% (corridor, about 2.6 m\u002Fs) where the wide-angle setup failed (Table II, Sec. VII-B)","Tolerates 2 s light outages by constant-velocity prediction corrected by lidar odometry (Sec. VII-B, Fig. 14)",[28,29,30,31,32],"Unsuitable for continuous darkness; the authors recommend lidar-only LOAM there (Sec. VII-B)","Fast motion blurs the point cloud and bends walls (Figs. 12 to 13)","The wide-angle camera loses feature tracking in fast turns, so motion estimation fails (Table II)","Scan matching fails in degenerate scenes dominated by planar areas (Sec. I)","(inference) Accuracy is measured from loop gaps, manually matched points or assumed flat walls, without an independent reference instrument (Sec. VII)",[34,35,36,37],"monocular camera (uEye monochrome at 60 Hz; wide-angle lens 76 deg or fisheye lens 185 deg horizontal FoV)","3D lidar built from a Hokuyo UTM-30LX 2D laser scanner rotated back-and-forth by a motor with encoder (1 s sweep)","KITTI configuration: single camera and Velodyne lidar","no IMU in the method or hardware",[39,40],"vehicle (KITTI odometry benchmark)","handheld (custom camera-lidar sensor carried by a person; about 0.7 m\u002Fs in Tests 1 and 2)","two sequential stages: frame-to-frame visual odometry solved by Levenberg-Marquardt in a robust-fitting framework, using features with depth from the lidar depthmap, depth from triangulation, or no depth; lidar odometry once per sweep with sweep-to-sweep refinement (linear drift model) and then sweep-to-map registration; transforms from both stages integrated into poses at image rate (Secs. IV to VI)","up to 300 Harris corners tracked by KLT over 5x6 image subregions; feature depth interpolated from the three nearest depthmap points found in a 2D KD-tree on angular coordinates; lidar edge and planar points selected by local curvature and matched to edge lines and planar patches with 3D KD-trees (sweep-to-sweep) or by eigenvalue analysis of local map clusters (sweep-to-map, ICP-style) (Secs. V to VII)","discrete frame-to-frame motion at the 60 Hz image rate; visual odometry drift modeled as linear (constant-velocity) motion within each 1 s sweep; lidar odometry at 1 Hz (Secs. I, IV, VI)","points registered with the visual odometry motion; residual distortion from visual drift removed by the linear-motion model in the sweep-to-sweep refinement (Sec. VI, Fig. 4)","none; the authors intentionally omit loop closure to focus on odometry (Sec. I)","none","incrementally built map point cloud: each distortion-free sweep is matched to the existing map cloud, with correspondences found by eigenvalue analysis of local point clusters, and then merged into it; edge and planar points of the previous sweep are kept in two 3D KD-trees for sweep-to-sweep matching (Sec. VI, Fig. 6)","none in the reported experiments; the authors state the method can be configured for localization only if a prior map is available (Sec. I)","registered 3D point cloud maps (Figs. 9 to 14) and 6-DoF poses at the image frame rate; export format not_reported","real-time on a laptop with 2.5 GHz quad cores under Linux; about 2.5 cores in total, 2 for visual odometry and 0.5 for lidar odometry (Sec. VII)",null,"not_verified",[],{"id":5,"kind":55,"shortName":7,"title":8,"authors":56,"year":9,"venue":59,"venueType":60,"publisher":61,"volumeIssuePages":62,"doi":63,"arxivId":51,"url":64,"firstPublicDate":65,"publicationStatus":16,"metadataStatus":66,"fulltextStatus":15,"era":10,"classicReason":67,"codeUrl":51,"cluster":11,"topics":68,"mdpi":69,"verification":70,"label":6,"fulltextRoute":71,"versionRead":72,"addedByCensus":69},"method",[57,58],"Ji Zhang","Sanjiv Singh","2015 IEEE International Conference on Robotics and Automation (ICRA)","conference","IEEE","pp. 2174-2181","10.1109\u002Ficra.2015.7139486","https:\u002F\u002Fapi.openalex.org\u002Fworks\u002Fdoi:10.1109\u002Ficra.2015.7139486","2015-05","metadata_verified","necessary technical node: one of the earliest visual-LiDAR odometry and mapping frameworks, coupling a monocular camera with a scanning 3D lidar without an IMU; visual odometry supplies high-rate motion and lidar scan matching refines motion and removes distortion (Secs. I, IV). Caution: later LIV papers (R3LIVE, SR-LIVO, Super Odometry) use the label 'V-LOAM' while citing the JFR 2018 paper (zhang2018lvio); the JFR paper itself benchmarks V-LOAM as a separate earlier method (JFR Table 4)",[11],false,"corrected","author copy","author-hosted PDF with the ICRA 2015 proceedings pagination (pp. 2174-2181) and IEEE copyright line, i.e., the published layout",[74,80,84,89,94,98,104,109],{"category":75,"model":76,"canonical":76,"role":77,"dataset":51,"specs":78,"locator":79},"camera","uEye monochrome camera","method input","60 Hz, 752 x 480 px, wide-angle lens with 76 deg horizontal FoV","Sec. VII, Sec. VII-A, Fig. 8",{"category":75,"model":81,"canonical":81,"role":77,"dataset":51,"specs":82,"locator":83},"second camera mounted underneath the original uEye camera, with fisheye lens","same configuration as the original camera except resolution 640 x 480 px; fisheye lens with 185 deg horizontal FoV; model not named separately","Sec. VII-A",{"category":85,"model":86,"canonical":86,"role":77,"dataset":51,"specs":87,"locator":88},"lidar","Hokuyo UTM-30LX","2D laser scanner, 180 deg FoV, 0.25 deg resolution, 40 lines\u002Fs; motor-actuated to form a 3D lidar","Sec. VII, Fig. 8",{"category":90,"model":91,"canonical":91,"role":77,"dataset":51,"specs":92,"locator":93},"other","motor and encoder actuating the scanner","rotates back-and-forth at 180 deg\u002Fs between -90 and 90 deg; encoder resolution 0.25 deg; one sweep lasts 1 s","Sec. IV, Sec. VII, Fig. 8",{"category":95,"model":96,"canonical":96,"role":77,"dataset":51,"specs":97,"locator":83},"platform","handheld custom camera-lidar sensor","carried by a person walking at about 0.7 m\u002Fs in the accuracy tests",{"category":99,"model":100,"canonical":100,"role":101,"dataset":51,"specs":102,"locator":103},"compute","laptop with 2.5 GHz quad cores","compute for runtime","Linux; method uses about 2.5 cores","Sec. VII",{"category":75,"model":105,"canonical":105,"role":106,"dataset":107,"specs":108,"locator":103},"single camera of the KITTI setup","dataset sensor","KITTI odometry benchmark","not_reported",{"category":85,"model":110,"canonical":110,"role":106,"dataset":107,"specs":108,"locator":103},"Velodyne lidar of the KITTI setup",[],{"totalRows":113,"groupCount":114,"groups":115,"others":291},28,6,[116,190,235,268],{"slug":117,"group":118,"sourceId":5,"sourceLabel":6,"table":119,"selfRows":120,"metrics":121,"seqs":125,"entrants":138,"cells":148,"outcomes":183,"locators":184,"hardware":186,"wordings":187,"notes":188},"vloam2015-table-i","vloam2015:Table I","Table I",16,[122],{"label":123,"unit":124,"statistic":108,"alignment":46},"Relative position error","% of distance travelled",[126,130,132,135],{"dataset":127,"sequence":128,"environment":129},"authors' custom camera-lidar sensor data","Test 1 (Loop 1), 49 m","indoor, handheld at 0.7 m\u002Fs; error from gap at loop closure",{"dataset":127,"sequence":131,"environment":129},"Test 1 (Loop 2), 47 m",{"dataset":127,"sequence":133,"environment":134},"Test 2, 186 m","outdoor, handheld; error from 15 manually matched lidar points at start and end",{"dataset":127,"sequence":136,"environment":137},"Test 3, 538 m","indoor and outdoor path through a building and two staircases; start-end position error",[139,142,144,146],{"name":140,"methodId":5,"linkable":141,"proposed":69,"self":141},"W-V (wide-angle camera, visual odometry only)",true,{"name":143,"methodId":5,"linkable":141,"proposed":69,"self":141},"F-V (fisheye camera, visual odometry only)",{"name":145,"methodId":5,"linkable":141,"proposed":141,"self":141},"W-VL (wide-angle camera, V-LOAM visual plus lidar odometry)",{"name":147,"methodId":5,"linkable":141,"proposed":141,"self":141},"F-VL (fisheye camera, V-LOAM visual plus lidar odometry)",[149,153,156,159,161,162,164,166,167,169,171,173,175,177,179,181],[150,150,150,151,152,150,152,152,150],0,1.1,-1,[154,150,150,155,152,150,152,152,150],1,1.8,[157,150,150,158,152,150,152,152,150],2,0.31,[160,150,150,158,152,150,152,152,150],3,[150,150,154,154,152,150,152,152,150],[154,150,154,163,152,150,152,152,150],2.1,[157,150,154,165,152,150,152,152,150],0.37,[160,150,154,165,152,150,152,152,150],[150,150,157,168,152,150,152,152,150],1.3,[154,150,157,170,152,150,152,152,150],2.7,[157,150,157,172,152,150,152,152,150],0.63,[160,150,157,174,152,150,152,152,150],0.64,[150,150,160,176,152,150,152,152,150],1.4,[154,150,160,178,152,150,152,152,150],3.1,[157,150,160,180,152,150,152,152,150],0.71,[160,150,160,182,152,150,152,152,150],0.73,[],[185],"Table I, Sec. VII-A",[],[],[189],"Relative position error as a fraction of distance travelled, based on 3D coordinates; no independent reference instrument",{"slug":191,"group":192,"sourceId":5,"sourceLabel":6,"table":193,"selfRows":194,"metrics":195,"seqs":197,"entrants":204,"cells":213,"outcomes":227,"locators":229,"hardware":231,"wordings":232,"notes":233},"vloam2015-table-ii","vloam2015:Table II","Table II",8,[196],{"label":123,"unit":124,"statistic":108,"alignment":46},[198,201],{"dataset":127,"sequence":199,"environment":200},"Test 4, 66 m","staircase with seven 180 deg turns; slow and fast trials; error from wall bending assuming flat aligned walls",{"dataset":127,"sequence":202,"environment":203},"Test 5, 54 m","corridor; slow and fast trials (about 2.6 m\u002Fs); error from gap at loop closure",[205,207,209,211],{"name":206,"methodId":5,"linkable":141,"proposed":141,"self":141},"V-LOAM W-S (wide-angle camera, slow motion)",{"name":208,"methodId":5,"linkable":141,"proposed":141,"self":141},"V-LOAM Fi-S (fisheye camera, slow motion)",{"name":210,"methodId":5,"linkable":141,"proposed":141,"self":141},"V-LOAM W-Fa (wide-angle camera, fast motion)",{"name":212,"methodId":5,"linkable":141,"proposed":141,"self":141},"V-LOAM Fi-Fa (fisheye camera, fast motion)",[214,216,218,219,220,222,224,225],[150,150,150,215,152,150,152,152,150],0.67,[154,150,150,217,152,150,152,152,150],0.68,[157,150,150,51,150,150,152,152,150],[160,150,150,168,152,150,152,152,150],[150,150,154,221,152,150,152,152,150],0.27,[154,150,154,223,152,150,152,152,150],0.28,[157,150,154,51,150,150,152,152,150],[160,150,154,226,152,150,152,152,150],0.39,[228],"failed",[230],"Table II, Sec. VII-B",[],[],[234],"Relative position errors in fast motion tests; 'Failed' = visual features lost tracking during fast turns",{"slug":236,"group":237,"sourceId":238,"sourceLabel":239,"table":240,"selfRows":154,"metrics":241,"seqs":247,"entrants":252,"cells":256,"outcomes":261,"locators":262,"hardware":264,"wordings":265,"notes":266},"sdvloam2023-table-vi-kitti-part","sdvloam2023:Table VI (KITTI part)","sdvloam2023","Yuan et al., 2023a","Table VI (KITTI part)",[242],{"label":243,"unit":244,"statistic":245,"alignment":246},"Relative translational error (RTE)","%","mean","not_applicable",[248],{"dataset":249,"sequence":250,"environment":251},"KITTI odometry","11-21 mean (KITTI test set)","urban, highway and country driving",[253,254],{"name":7,"methodId":5,"linkable":141,"proposed":69,"self":141},{"name":255,"methodId":238,"linkable":141,"proposed":141,"self":69},"Ours (SDV-LOAM)",[257,259],[150,150,150,258,152,150,152,152,150],0.54,[154,150,150,260,152,150,152,152,150],0.6,[],[263],"Table VI; Sec. VII",[],[],[267],"KITTI odometry; relative translational error (%) of visual-LiDAR odometry; '+' = open-source LiDAR odometry modified by the authors to use the SDV-LOAM visual module as motion prior; V-LOAM only has test-set results; '-' cells not stored",{"slug":269,"group":270,"sourceId":5,"sourceLabel":6,"table":271,"selfRows":154,"metrics":272,"seqs":275,"entrants":279,"cells":281,"outcomes":284,"locators":285,"hardware":287,"wordings":288,"notes":289},"vloam2015-text-sec-viii","vloam2015:Text Sec. VIII","Text Sec. VIII",[273],{"label":274,"unit":244,"statistic":245,"alignment":108},"relative position drift (benchmark ranking by average translation and rotation errors)",[276],{"dataset":107,"sequence":277,"environment":278},"benchmark average","vehicle-mounted single camera and Velodyne lidar",[280],{"name":7,"methodId":5,"linkable":141,"proposed":141,"self":141},[282],[150,150,150,283,152,150,152,152,150],0.75,[],[286],"Abstract, Sec. VIII",[],[],[290],"KITTI odometry benchmark result as stated by the authors (ranked first at the time); no per-sequence table in the paper",[292,299],{"group":293,"slug":294,"sourceLabel":295,"table":296,"selfRows":154,"datasets":297},"zhang2016degeneracy:Text Sec.VI","zhang2016degeneracy-text-sec-vi","Zhang et al., 2016","Text Sec.VI",[298],"authors' own data (Test 3)",{"group":300,"slug":301,"sourceLabel":302,"table":303,"selfRows":154,"datasets":304},"zhang2018lvio:Table 4","zhang2018lvio-table-4","Zhang & Singh, 2018","Table 4",[305],"authors' data, Accuracy Test 2",1790510660542]