[{"data":1,"prerenderedAt":436},["ShallowReactive",2],{"method-maplab2_2023":3},{"method":4,"reference":59,"equipment":86,"figures":125,"results":126},{"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":28,"sensors":34,"platform":40,"estimator":43,"association":44,"timeModel":45,"deskew":46,"loopClosure":47,"globalOptimization":48,"mapRepresentation":49,"prior":50,"outputGeometry":51,"compute":52,"codeUrl":53,"codeLicense":54,"relatedVersions":55},"maplab2_2023","Cramariuc et al., 2023","maplab 2.0","maplab 2.0 - A Modular and Multi-Modal Mapping Framework",2023,"recent","C06","estimation_framework_or_library","maplab 2.0 是以因子圖為核心的模組化、多模態建圖框架：一張地圖由多個任務（mission，即單次連續建圖時段）組成，頂點包含位姿、速度、IMU 偏差與地標，可整合視覺、光達與語意地標。新版加入 mapping server，把各機器人的子地圖先各自做局部最佳化與迴圈，再於全域層級做跨機迴圈與聯合最佳化；離線 console 提供批次最佳化、地圖合併、以 ICP\u002FG-ICP 產生並以可切換約束抗離群的光達迴圈，以及 Voxblox 稠密重建。","maplab 2.0 is an open factor-graph mapping framework supporting multi-modal sensors, multi-session and multi-robot map merging with visual and LiDAR loop closures, and offline map processing.","full_text_reviewed","peer_reviewed_published","main_body","使用 Hilti 2021 SLAM 資料集，作者描述其含室內辦公與戶外工地場景（Sec. IV-A）。Table II 中 Construction 1 與 Construction 2 的 APE RMSE，maplab 2.0（FAST-LIO2 + SP + B）為 0.04 m 與 0.07 m，與單獨 FAST-LIO2 相同，明顯改善只出現在 Parking 序列（5.00 m 降至 0.21 m）。大型訓練設施含倒塌建築與狹窄空間的 23 趟手持多時段建圖僅有定性展示（Sec. IV-B、Fig. 4）。",[20,21,22],"public_benchmark","real_construction_site","independent_reference",[24,25,26,27],"Refines the best odometry: FAST-LIO2 + SP + B reduced Parking APE RMSE from 5.00 m (FAST-LIO2) to 0.21 m and matched or improved FAST-LIO2 on the other eight Hilti sequences (Table II)","Among vision-based methods, maplab 2.0 configurations outperform ORB-SLAM3, RTAB-Map, maplab and ROVIO\u002FOKVIS odometry on most sequences (Table II, Sec. IV-A)","Handheld 23-run multi-session mapping over about two hours and 10 km in a training facility, merged with visual loops and optional RTK constraints (Sec. IV-B)","Server-based multi-robot processing matched sequential multi-session accuracy on EuRoC (0.043 m average APE RMSE) while running about ten times faster (Sec. IV-B)",[29,30,31,32,33],"Loop-closure edges from LiDAR registration and visual matching use predefined, empirically chosen fixed covariances (Sec. III-A, III-D, IV-A); only the semantic-object loop-closure demo computes the constraint covariance (Sec. IV-D)","(inference) LiDAR loop closure uses pairwise ICP\u002FG-ICP; no BA-style point-cloud consistency refinement is described in the full text","(inference) Every maplab 2.0 configuration needed more total time than the 52 min dataset duration: 98 to 267 min, and 194 min for the most accurate FAST-LIO2 + SP + B, on i7-8700 with RTX 2080 (Table II)","LiDAR-image keypoint landmarks suffer outliers from missing points and moving objects (Sec. IV-C)","On the two construction sequences the best maplab 2.0 result equals FAST-LIO2 alone (0.04 m and 0.07 m APE RMSE) (Table II); (inference) no construction-specific accuracy gain is shown",[35,36,37,38,39],"3D LiDAR","IMU (optional; framework does not require an IMU)","multi-camera or monocular camera","GNSS (RTK, optional absolute constraints)","wheel encoders and RGB-D landmarks (supported interfaces, not evaluated)",[41,42],"handheld","UAV (EuRoC MAV benchmark)","factor graph over vertices (6-DoF pose, velocity, IMU biases, landmarks) with batch optimization \u002F bundle adjustment","visual landmarks from ORB detection with BRISK or FREAK binary descriptors (inverted multi-index matching), plus external float descriptors such as SuperPoint with SuperGlue tracking and SIFT with Lucas-Kanade tracking (PCA-compressed 256 to 32, FLANN matching); 2D-3D matches with covisibility filtering and P3P in RANSAC for visual loop closure, or landmark merging; 3D landmarks (RGB-D, LiDAR image keypoints) matched by 3D-3D RANSAC; LiDAR loop closures by ICP or G-ICP registration in the console","discrete poses","not_reported (delegated to odometry source, e.g., FAST-LIO2)","visual and LiDAR intra- and inter-mission loop closures; loop edges as switchable constraints","global multi-mission, multi-robot optimization in mapping server or offline console","factor-graph map of missions with attached sensor data; dense reconstruction via Voxblox plugin","optional RTK GNSS absolute pose constraints","optimized poses, landmarks; dense volumetric reconstruction via Voxblox plugin; map data export (Sec. III-D)","online mapping node and server plus offline console; Table II timings on Intel i7-8700 with Nvidia RTX 2080 GPU: 52 min of Hilti data processed in 98 to 267 min by the maplab 2.0 configurations (194 min for FAST-LIO2 + SP + B) versus 52 min for FAST-LIO2 alone; EuRoC server run 3 min 27 s versus 35 min 56 s sequentially (Sec. IV-B)","https:\u002F\u002Fgithub.com\u002Fethz-asl\u002Fmaplab","Apache-2.0 (LICENSE file)",[56],{"relation":57,"title":58,"doi_or_url":53},"code_release","ethz-asl\u002Fmaplab",{"id":5,"kind":60,"shortName":7,"title":61,"authors":62,"year":9,"venue":71,"venueType":72,"publisher":73,"volumeIssuePages":74,"doi":75,"arxivId":76,"url":77,"firstPublicDate":78,"publicationStatus":16,"metadataStatus":79,"fulltextStatus":15,"era":10,"classicReason":80,"codeUrl":53,"cluster":11,"topics":81,"mdpi":82,"verification":83,"label":6,"fulltextRoute":84,"versionRead":85,"addedByCensus":82},"software","maplab 2.0 – A Modular and Multi-Modal Mapping Framework",[63,64,65,66,67,68,69,70],"Andrei Cramariuc","Lukas Bernreiter","Florian Tschopp","Marius Fehr","Victor Reijgwart","Juan Nieto","Roland Siegwart","Cesar Cadena","IEEE Robotics and Automation Letters","journal","IEEE","8(2):520-527","10.1109\u002Flra.2022.3227865","2212.00654","https:\u002F\u002Farxiv.org\u002Fabs\u002F2212.00654","2022-12-01","metadata_verified","not_applicable",[11],false,"corrected","arXiv","arXiv 2212.00654v2 (3 Jan 2023), header 'IEEE RA-L preprint version, accepted November 2022'; version of record not compared",[87,94,98,102,109,111,115,119,123],{"category":88,"model":89,"canonical":89,"role":90,"dataset":91,"specs":92,"locator":93},"camera","five cameras (Hilti 2021 rig; model not stated in paper)","dataset sensor","HILTI 2021 SLAM Dataset","five cameras; frontal camera or stereo pair used for ROVIO\u002FOKVIS odometry, all five for loop closure","Sec. IV-A",{"category":95,"model":96,"canonical":96,"role":90,"dataset":91,"specs":97,"locator":93},"imu","ADIS IMU","not_reported",{"category":99,"model":100,"canonical":101,"role":90,"dataset":91,"specs":97,"locator":93},"lidar","OS0-64","Ouster OS0-64",{"category":103,"model":104,"canonical":104,"role":105,"dataset":106,"specs":107,"locator":108},"mobile_scanner_device","handheld device with five cameras and an Ouster OS0-128","method input",null,"23 runs, more than two hours, about 10 km, indoor-outdoor transitions; OKVIS odometry","Sec. IV-B",{"category":99,"model":110,"canonical":110,"role":105,"dataset":106,"specs":97,"locator":108},"Ouster OS0-128",{"category":112,"model":113,"canonical":113,"role":105,"dataset":106,"specs":114,"locator":108},"gnss","RTK GPS","optional absolute pose constraints where available",{"category":88,"model":116,"canonical":116,"role":105,"dataset":117,"specs":97,"locator":118},"RGB-inertial sensor (cited as VersaVIS, ref. [52])","custom indoor office semantic dataset","Sec. IV-D",{"category":120,"model":121,"canonical":121,"role":122,"dataset":106,"specs":97,"locator":93},"compute","Intel i7-8700","compute for runtime",{"category":120,"model":124,"canonical":124,"role":122,"dataset":106,"specs":97,"locator":93},"Nvidia RTX 2080 GPU",[],{"totalRows":127,"groupCount":128,"groups":129,"others":435},44,2,[130,401],{"slug":131,"group":132,"sourceId":5,"sourceLabel":6,"table":133,"selfRows":134,"metrics":135,"seqs":143,"entrants":170,"cells":199,"outcomes":394,"locators":395,"hardware":396,"wordings":398,"notes":399},"maplab2-2023-table-ii","maplab2_2023:Table II","Table II",40,[136,140],{"label":137,"unit":138,"statistic":139,"alignment":97},"RMSE of the absolute position error (APE)","m","RMSE",{"label":141,"unit":142,"statistic":97,"alignment":80},"Total Time (processing all sequences)","min",[144,147,149,152,154,157,159,162,165,167],{"dataset":91,"sequence":145,"environment":146},"Construction 1","outdoor construction site (dataset description)",{"dataset":91,"sequence":148,"environment":146},"Construction 2",{"dataset":91,"sequence":150,"environment":151},"IC Office","indoor office",{"dataset":91,"sequence":153,"environment":151},"Office Mitte",{"dataset":91,"sequence":155,"environment":156},"Basement 3","basement (per sequence name)",{"dataset":91,"sequence":158,"environment":156},"Basement 4",{"dataset":91,"sequence":160,"environment":161},"Parking","parking (per sequence name)",{"dataset":91,"sequence":163,"environment":164},"Campus 1","campus (per sequence name)",{"dataset":91,"sequence":166,"environment":164},"Campus 2",{"dataset":91,"sequence":168,"environment":169},"all 9 sequences (total dataset duration 52 min)","mixed indoor and outdoor",[171,175,178,181,183,185,188,191,193,195,197],{"name":172,"methodId":173,"linkable":174,"proposed":82,"self":82},"ORB SLAM3","orbslam3_2021",true,{"name":176,"methodId":177,"linkable":174,"proposed":82,"self":82},"LVI SAM","lvisam2021",{"name":179,"methodId":180,"linkable":174,"proposed":82,"self":82},"RTAB Map","rtabmap2019",{"name":182,"methodId":106,"linkable":82,"proposed":82,"self":82},"maplab",{"name":184,"methodId":106,"linkable":82,"proposed":82,"self":82},"ROVIO",{"name":186,"methodId":187,"linkable":174,"proposed":82,"self":82},"OKVIS","okvis2015",{"name":189,"methodId":190,"linkable":174,"proposed":82,"self":82},"FAST LIO2","fastlio2_2022",{"name":192,"methodId":5,"linkable":174,"proposed":174,"self":174},"maplab 2.0: ROVIO + SIFT",{"name":194,"methodId":5,"linkable":174,"proposed":174,"self":174},"maplab 2.0: OKVIS + SP + B",{"name":196,"methodId":5,"linkable":174,"proposed":174,"self":174},"maplab 2.0: OKVIS + SP + B + ICP",{"name":198,"methodId":5,"linkable":174,"proposed":174,"self":174},"maplab 2.0: FAST-LIO2 + SP + B",[200,204,207,209,212,215,218,221,224,227,229,231,233,235,237,239,241,243,245,247,249,250,251,253,255,256,258,260,262,263,264,265,266,267,269,271,273,275,277,279,280,282,284,286,288,289,290,292,293,295,297,299,300,301,302,303,305,306,307,309,311,313,314,315,316,317,318,320,322,324,326,328,330,331,333,335,336,337,339,340,342,344,346,348,349,350,351,353,354,356,357,359,361,363,365,366,368,370,371,372,374,376,378,380,382,384,386,388,390,392],[201,201,201,202,203,201,203,203,201],0,1.55,-1,[205,201,201,206,203,201,203,203,201],1,0.13,[128,201,201,208,203,201,203,203,201],0.36,[210,201,201,211,203,201,203,203,201],3,0.16,[213,201,201,214,203,201,203,203,201],4,0.98,[216,201,201,217,203,201,203,203,201],5,1.17,[219,201,201,220,203,201,203,203,201],6,0.04,[222,201,201,223,203,201,203,203,201],7,0.14,[225,201,201,226,203,201,203,203,201],8,0.08,[228,201,201,226,203,201,203,203,201],9,[230,201,201,220,203,201,203,203,201],10,[201,201,205,232,203,201,203,203,201],2.77,[205,201,205,234,203,201,203,203,201],0.33,[128,201,205,236,203,201,203,203,201],0.67,[210,201,205,238,203,201,203,203,201],0.57,[213,201,205,240,203,201,203,203,201],1.5,[216,201,205,242,203,201,203,203,201],2.13,[219,201,205,244,203,201,203,203,201],0.07,[222,201,205,246,203,201,203,203,201],0.34,[225,201,205,248,203,201,203,203,201],0.19,[228,201,205,248,203,201,203,203,201],[230,201,205,244,203,201,203,203,201],[201,201,128,252,203,201,203,203,201],1.86,[205,201,128,254,203,201,203,203,201],0.12,[128,201,128,240,203,201,203,203,201],[210,201,128,257,203,201,203,203,201],0.09,[213,201,128,259,203,201,203,203,201],1.16,[216,201,128,261,203,201,203,203,201],1.27,[219,201,128,226,203,201,203,203,201],[222,201,128,226,203,201,203,203,201],[225,201,128,226,203,201,203,203,201],[228,201,128,244,203,201,203,203,201],[230,201,128,244,203,201,203,203,201],[201,201,210,268,203,201,203,203,201],1.7,[205,201,210,270,203,201,203,203,201],0.24,[128,201,210,272,203,201,203,203,201],0.94,[210,201,210,274,203,201,203,203,201],3.18,[213,201,210,276,203,201,203,203,201],0.86,[216,201,210,278,203,201,203,203,201],1.15,[219,201,210,254,203,201,203,203,201],[222,201,210,281,203,201,203,203,201],0.27,[225,201,210,283,203,201,203,203,201],0.18,[228,201,210,285,203,201,203,203,201],0.15,[230,201,210,287,203,201,203,203,201],0.1,[201,201,213,202,203,201,203,203,201],[205,201,213,287,203,201,203,203,201],[128,201,213,291,203,201,203,203,201],0.38,[210,201,213,257,203,201,203,203,201],[213,201,213,294,203,201,203,203,201],3.05,[216,201,213,296,203,201,203,203,201],1.01,[219,201,213,298,203,201,203,203,201],0.05,[222,201,213,257,203,201,203,203,201],[225,201,213,257,203,201,203,203,201],[228,201,213,226,203,201,203,203,201],[230,201,213,298,203,201,203,203,201],[201,201,216,304,203,201,203,203,201],1.71,[205,201,216,206,203,201,203,203,201],[128,201,216,291,203,201,203,203,201],[210,201,216,308,203,201,203,203,201],0.11,[213,201,216,310,203,201,203,203,201],2.9,[216,201,216,312,203,201,203,203,201],1.23,[219,201,216,220,203,201,203,203,201],[222,201,216,308,203,201,203,203,201],[225,201,216,287,203,201,203,203,201],[228,201,216,257,203,201,203,203,201],[230,201,216,220,203,201,203,203,201],[201,201,219,319,203,201,203,203,201],5.49,[205,201,219,321,203,201,203,203,201],4.43,[128,201,219,323,203,201,203,203,201],7.82,[210,201,219,325,203,201,203,203,201],0.39,[213,201,219,327,203,201,203,203,201],6.13,[216,201,219,329,203,201,203,203,201],3.36,[219,201,219,216,203,201,203,203,201],[222,201,219,332,203,201,203,203,201],0.31,[225,201,219,334,203,201,203,203,201],0.21,[228,201,219,334,203,201,203,203,201],[230,201,219,334,203,201,203,203,201],[201,201,222,338,203,201,203,203,201],1.93,[205,201,222,254,203,201,203,203,201],[128,201,222,341,203,201,203,203,201],0.93,[210,201,222,343,203,201,203,203,201],0.6,[213,201,222,345,203,201,203,203,201],5.1,[216,201,222,347,203,201,203,203,201],2.41,[219,201,222,244,203,201,203,203,201],[222,201,222,291,203,201,203,203,201],[225,201,222,248,203,201,203,203,201],[228,201,222,352,203,201,203,203,201],0.17,[230,201,222,244,203,201,203,203,201],[201,201,225,355,203,201,203,203,201],2.24,[205,201,225,223,203,201,203,203,201],[128,201,225,358,203,201,203,203,201],0.79,[210,201,225,360,203,201,203,203,201],0.47,[213,201,225,362,203,201,203,203,201],2.02,[216,201,225,364,203,201,203,203,201],2.23,[219,201,225,257,203,201,203,203,201],[222,201,225,367,203,201,203,203,201],0.28,[225,201,225,369,203,201,203,203,201],0.2,[228,201,225,283,203,201,203,203,201],[230,201,225,226,203,201,203,203,201],[201,205,228,373,203,201,201,203,201],61,[205,205,228,375,203,201,201,203,201],68,[128,205,228,377,203,201,201,203,201],163,[210,205,228,379,203,201,201,203,201],82,[213,205,228,381,203,201,201,203,201],58,[216,205,228,383,203,201,201,203,201],121,[219,205,228,385,203,201,201,203,201],52,[222,205,228,387,203,201,201,203,201],98,[225,205,228,389,203,201,201,203,201],236,[228,205,228,391,203,201,201,203,201],267,[230,205,228,393,203,201,201,203,201],194,[],[133],[397],"Intel i7-8700 + Nvidia RTX 2080 GPU",[],[400],"HILTI 2021 SLAM Dataset, RMSE of APE; seven baselines and four maplab 2.0 configurations (ROVIO + SIFT, OKVIS + SP + B, OKVIS + SP + B + ICP, FAST-LIO2 + SP + B); sensors used differ per method (icons in table); Total Time measured on Intel i7-8700 with Nvidia RTX 2080 for 52 min of data",{"slug":402,"group":403,"sourceId":5,"sourceLabel":6,"table":404,"selfRows":213,"metrics":405,"seqs":411,"entrants":416,"cells":421,"outcomes":429,"locators":430,"hardware":431,"wordings":432,"notes":433},"maplab2-2023-text-sec-iv-b","maplab2_2023:Text Sec.IV-B","Text Sec.IV-B",[406,408],{"label":407,"unit":138,"statistic":139,"alignment":97},"average RMSE APE",{"label":409,"unit":410,"statistic":97,"alignment":80},"total time incl. odometry, optimization and map merging (3 min 27 s vs 35 min 56 s)","s",[412],{"dataset":413,"sequence":414,"environment":415},"EuRoC MAV","all 11 sequences","not_reported (EuRoC MAV benchmark)",[417,419],{"name":418,"methodId":5,"linkable":174,"proposed":174,"self":174},"maplab 2.0 mapping server (parallel, 11 missions)",{"name":420,"methodId":5,"linkable":174,"proposed":174,"self":174},"maplab 2.0 sequential multi-session (mapping node + console)",[422,424,426,427],[201,201,201,423,203,201,203,203,201],0.043,[201,205,201,425,203,201,203,203,201],207,[205,201,201,423,203,201,203,203,201],[205,205,201,428,203,201,203,203,201],2156,[],[108],[],[],[434],"EuRoC 11 sequences with ROVIO and BRISK: parallel multi-robot mapping server vs sequential multi-session console workflow",[],1790510657354]