[{"data":1,"prerenderedAt":496},["ShallowReactive",2],{"method-zhang1994icp":3},{"method":4,"reference":56,"equipment":74,"figures":96,"results":97},{"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":26,"sensors":33,"platform":37,"estimator":40,"association":41,"timeModel":42,"deskew":43,"loopClosure":44,"globalOptimization":44,"mapRepresentation":45,"prior":46,"outputGeometry":47,"compute":48,"codeUrl":49,"codeLicense":50,"relatedVersions":51},"zhang1994icp","Zhang, 1994","Iterative point matching (Zhang)","Iterative point matching for registration of free-form curves and surfaces",1994,"classic","C02","registration_component","本文提出迭代虛擬點匹配（iterative pseudo point matching）演算法，用於配準邊緣式立體視覺取得的三維曲線，或相關式立體視覺重建的稠密三維地圖。方法假設兩次觀測之間的運動很小，或已由里程計與慣性系統近似得知；每次迭代先以目前估計轉換第一組點，再以 k-D tree 搜尋第二組資料中的最近點，接著依配對距離的平均值與標準差動態設定最大容許距離 Dmax，剔除離群點、遮蔽以及出現或消失的點（曲線另以切線夾角不超過 60 度限制配對），最後以四元數或對偶四元數最小平方法求剛體運動並反覆至收斂。作者以合成曲線、移動載具上三目立體相機拍攝的椅子場景、相關式立體視覺重建的岩石場景以及頭像距離影像驗證。","Iterative pseudo point matching: closest-point pairing with a k-D tree, an adaptive distance threshold Dmax set each iteration from the mean and standard deviation of pair distances to reject outliers, occlusion and appearance or disappearance, and closed-form (dual) quaternion least-squares rigid motion; assumes small or approximately known motion.","full_text_reviewed","peer_reviewed_published","background","not_reported (no construction use; application context is vision-based autonomous vehicle navigation in rugged terrain and Digital Elevation Map building, Sec. 1)",[],[21,22,23,24,25],"Handles gross outliers, appearance and disappearance, and occlusion through dynamic distance statistics (Sec. 3.3, Sec. 8, Sec. 9)","Simple, extensible and general point-set representation for arbitrary shapes (Sec. 9)","Efficient: closest-point search O(N log N); coarse-to-fine sampling cut runtime from 7.49 s to 3.39 s on synthetic curves and from 32.5 s to 10.5 s on the chair scene with little accuracy change (Sec. 5.2, Sec. 5.3)","No smoothing or derivative estimation required (Sec. 9)","Rock scene: 0.65 to 0.92 deg and 2.87 to 6.18 cm difference from the manual registration from initial offsets of 20 deg with 2.07 to 2.56 m (Test 2 needed 80 iterations) (Sec. 6.1)",[27,28,29,30,31,32],"Converges only to the closest local minimum; not appropriate for large motion (Sec. 7.6, Sec. 9)","User parameter D must be set and affects the convergence rate (Sec. 4.1, Sec. 9)","Only partially accounts for measurement uncertainty; a full treatment would need Kalman filtering at higher cost (Sec. 7.5, Sec. 9)","Works better on rugged terrain than on flat ground, where many close local minima exist (Sec. 9)","Monotonic convergence is not guaranteed because p_i switches between 0 and 1 (Sec. 7.2)","Closest sample point approximation makes the result depend on sampling density (Sec. 7.4)",[34,35,36],"trinocular edge-based stereo (3-D curves)","correlation-based stereo (dense 3-D maps)","range images (head figure, Sec. 6.2; sensor not reported)",[38,39],"mobile vehicle carrying a trinocular stereo system (Sec. 5.3)","synthetic data (Sec. 5.1-5.2)","Closed-form least-squares rigid motion from the retained pairs (quaternion method and dual number quaternion method of Walker et al. 1991, both implemented with identical results), iterated until the relative changes of r and t are both below 1% or a maximum of 20 (curves) or 40 (surfaces) iterations is reached","Closest sample point in the second frame found with a 3-D tree whose search radius shrinks with Dmax; a pair is removed when its distance exceeds an adaptive Dmax computed each iteration from the mean mu and standard deviation sigma of pair distances (mu+3sigma, mu+2sigma, mu+sigma, or the histogram valley after the main peak, depending on mu relative to the user parameter D); for curves, pairs whose tangent angle exceeds 60 deg are also rejected","not_applicable (pairwise rigid registration)","not_applicable","none","Raw 3-D point sets: chained points for curves and scattered points for dense 3-D maps, used without smoothing or primitive fitting","Small inter-frame motion or an approximate initial motion (e.g., from odometric and inertial systems) is required; the algorithm converges to the closest local minimum, and large motions need a global method first or sampling of the 6-D motion space","6-DoF rigid transformation","C implementation (not optimized) run on a SUN 4\u002F60 workstation (double precision LINPACK 1.05 Mflops); per-iteration complexity O(m log n) with the k-D tree; e.g., 6.5 s for the 200-point synthetic curve case (15 iterations) and 32.5 s for the real chair scene (12 iterations), reduced to 10.5 s with coarse-to-fine sampling",null,"not_verified",[52],{"relation":53,"title":54,"doi_or_url":55},"preprint","INRIA Research Report RR-1658 (1992), 'Iterative point matching for registration of free-form curves' (report title covers curves only; content equivalence with the IJCV article not verified)","https:\u002F\u002Finria.hal.science\u002Finria-00074899",{"id":5,"kind":57,"shortName":7,"title":8,"authors":58,"year":9,"venue":60,"venueType":61,"publisher":62,"volumeIssuePages":63,"doi":64,"arxivId":49,"url":65,"firstPublicDate":66,"publicationStatus":16,"metadataStatus":67,"fulltextStatus":15,"era":10,"classicReason":68,"codeUrl":49,"cluster":11,"topics":69,"mdpi":70,"verification":71,"label":6,"fulltextRoute":72,"versionRead":73,"addedByCensus":70},"method",[59],"Zhengyou Zhang","International Journal of Computer Vision","journal","Springer","13(2):119-152","10.1007\u002Fbf01427149","https:\u002F\u002Fapi.crossref.org\u002Fworks\u002F10.1007\u002FBF01427149","1994-10","metadata_verified","necessary technical node: independent near-simultaneous closest-point iterative registration (the paper itself names Besl and McKay, Chen and Medioni, Menq et al. and Champleboux et al. as independent similar work) that adds an adaptive statistics-based distance threshold for outlier, occlusion and appearance handling plus k-D tree closest-point search; confirmed from the primary text.",[11],false,"corrected","NTU institutional (curl)","Version of record: International Journal of Computer Vision 13(2):119-152 (1994), Kluwer Academic Publishers; scanned publisher PDF with OCR text layer (34 pages incl. colour figure section). Tables 2, 3 and 4 were checked against rendered page images (pp. 132, 133, 140) because the OCR of Table 4 is garbled.",[75,81,85,90],{"category":76,"model":77,"canonical":77,"role":78,"dataset":49,"specs":79,"locator":80},"stereo_camera","trinocular stereo system","method input","mounted on the authors' mobile vehicle; 3-D curves reconstructed with the curve-based trinocular stereo algorithm of Robert and Faugeras (1991); in the chair test the chair was about 3 m away and the two positions differed by about 4 deg and 100 mm (36 curves and 588 points, 48 curves and 763 points); camera models and optics not reported","Sec. 5.3; Fig. 17",{"category":76,"model":82,"canonical":82,"role":78,"dataset":49,"specs":83,"locator":84},"stereo rig with a correlation-based stereovision system (image triplets)","about 6 m from the rock scene; 71505 and 51503 reconstructed points; positions differ by 30 deg and 3.75 m; data resolution about 5 cm","Sec. 6.1; Figs. 19-20; Sec. 7.2",{"category":86,"model":87,"canonical":87,"role":78,"dataset":49,"specs":88,"locator":89},"platform","mobile vehicle (unnamed)","carries the trinocular stereo system; context of autonomous navigation in rugged terrain","Sec. 1; Sec. 5.3",{"category":91,"model":92,"canonical":92,"role":93,"dataset":49,"specs":94,"locator":95},"compute","SUN 4\u002F60 workstation","compute for runtime","double precision LINPACK rating 1.05 Mflops; C implementation, not optimized","Sec. 5 opening; Note 2",[],{"totalRows":98,"groupCount":99,"groups":100,"others":462},127,10,[101,223,331,425],{"slug":102,"group":103,"sourceId":5,"sourceLabel":6,"table":104,"selfRows":105,"metrics":106,"seqs":116,"entrants":131,"cells":137,"outcomes":217,"locators":218,"hardware":219,"wordings":220,"notes":221},"zhang1994icp-table-4","zhang1994icp:Table 4","Table 4",36,[107,111,113],{"label":108,"unit":109,"statistic":110,"alignment":44},"rotation error (%)","%","mean",{"label":112,"unit":109,"statistic":110,"alignment":44},"translation error (%)",{"label":114,"unit":115,"statistic":110,"alignment":44},"execution time (unit not restated in Table 4; all other times in the paper are in seconds)","s",[117,121,123,125,127,129],{"dataset":118,"sequence":119,"environment":120},"synthetic 3-D curve (Sec. 5.1)","noise std 0","synthetic",{"dataset":118,"sequence":122,"environment":120},"noise std 2",{"dataset":118,"sequence":124,"environment":120},"noise std 4",{"dataset":118,"sequence":126,"environment":120},"noise std 6",{"dataset":118,"sequence":128,"environment":120},"noise std 8",{"dataset":118,"sequence":130,"environment":120},"noise std 10",[132,135],{"name":133,"methodId":5,"linkable":134,"proposed":134,"self":134},"non-symmetric criterion (2), default",true,{"name":136,"methodId":5,"linkable":134,"proposed":134,"self":134},"symmetric criterion (1), variant",[138,142,145,147,149,152,154,157,159,162,164,167,169,171,173,175,177,179,181,183,185,187,189,191,193,195,197,199,201,203,205,207,209,211,213,215],[139,139,139,140,141,139,141,141,139],0,1.81,-1,[143,139,139,144,141,139,141,141,139],1,0.12,[139,139,143,146,141,139,141,141,139],4.36,[143,139,143,148,141,139,141,141,139],2.68,[139,139,150,151,141,139,141,141,139],2,4.6,[143,139,150,153,141,139,141,141,139],3.63,[139,139,155,156,141,139,141,141,139],3,7.56,[143,139,155,158,141,139,141,141,139],6.4,[139,139,160,161,141,139,141,141,139],4,11.35,[143,139,160,163,141,139,141,141,139],8.52,[139,139,165,166,141,139,141,141,139],5,11.94,[143,139,165,168,141,139,141,141,139],8.36,[139,143,139,170,141,139,141,141,139],8.22,[143,143,139,172,141,139,141,141,139],5.83,[139,143,143,174,141,139,141,141,139],7.38,[143,143,143,176,141,139,141,141,139],6.35,[139,143,150,178,141,139,141,141,139],8.56,[143,143,150,180,141,139,141,141,139],7.32,[139,143,155,182,141,139,141,141,139],7.61,[143,143,155,184,141,139,141,141,139],6.42,[139,143,160,186,141,139,141,141,139],7.36,[143,143,160,188,141,139,141,141,139],7.08,[139,143,165,190,141,139,141,141,139],8.96,[143,143,165,192,141,139,141,141,139],7.92,[139,150,139,194,141,139,139,141,139],5.04,[143,150,139,196,141,139,139,141,139],10.04,[139,150,143,198,141,139,139,141,139],5.32,[143,150,143,200,141,139,139,141,139],10.8,[139,150,150,202,141,139,139,141,139],6.1,[143,150,150,204,141,139,139,141,139],12.54,[139,150,155,206,141,139,139,141,139],7.64,[143,150,155,208,141,139,139,141,139],15.88,[139,150,160,210,141,139,139,141,139],8.18,[143,150,160,212,141,139,139,141,139],16.74,[139,150,165,214,141,139,139,141,139],9.98,[143,150,165,216,141,139,139,141,139],20.48,[],[104],[92],[],[222],"Same synthetic curve; non-symmetric criterion (2) vs symmetric criterion (1); 10 iterations; mean of 10 tries; execution-time unit not restated in Table 4 (seconds in Tables 2-3)",{"slug":224,"group":225,"sourceId":5,"sourceLabel":6,"table":226,"selfRows":227,"metrics":228,"seqs":235,"entrants":252,"cells":255,"outcomes":325,"locators":326,"hardware":327,"wordings":328,"notes":329},"zhang1994icp-table-2","zhang1994icp:Table 2","Table 2",33,[229,231,233],{"label":230,"unit":109,"statistic":110,"alignment":44},"rotation error ||r - r_est||\u002F||r|| (%)",{"label":232,"unit":109,"statistic":110,"alignment":44},"translation error ||t - t_est||\u002F||t|| (%)",{"label":234,"unit":115,"statistic":110,"alignment":44},"execution time (s) per two-frame registration",[236,237,238,239,240,241,242,244,246,248,250],{"dataset":118,"sequence":119,"environment":120},{"dataset":118,"sequence":122,"environment":120},{"dataset":118,"sequence":124,"environment":120},{"dataset":118,"sequence":126,"environment":120},{"dataset":118,"sequence":128,"environment":120},{"dataset":118,"sequence":130,"environment":120},{"dataset":118,"sequence":243,"environment":120},"noise std 12",{"dataset":118,"sequence":245,"environment":120},"noise std 14",{"dataset":118,"sequence":247,"environment":120},"noise std 16",{"dataset":118,"sequence":249,"environment":120},"noise std 18",{"dataset":118,"sequence":251,"environment":120},"noise std 20",[253],{"name":254,"methodId":5,"linkable":134,"proposed":134,"self":134},"iterative pseudo point matching (proposed)",[256,258,260,262,264,266,268,271,274,277,280,282,284,285,287,289,291,293,295,297,299,301,303,305,307,309,311,313,315,317,319,321,323],[139,139,139,257,141,139,141,141,139],2.25,[139,139,143,259,141,139,141,141,139],2.12,[139,139,150,261,141,139,141,141,139],4.63,[139,139,155,263,141,139,141,141,139],9.62,[139,139,160,265,141,139,141,141,139],13.73,[139,139,165,267,141,139,141,141,139],14.31,[139,139,269,270,141,139,141,141,139],6,20.47,[139,139,272,273,141,139,141,141,139],7,18.07,[139,139,275,276,141,139,141,141,139],8,23.87,[139,139,278,279,141,139,141,141,139],9,37.04,[139,139,99,281,141,139,141,141,139],33.2,[139,143,139,283,141,139,141,141,139],1.77,[139,143,143,146,141,139,141,141,139],[139,143,150,286,141,139,141,141,139],4.55,[139,143,155,288,141,139,141,141,139],4.84,[139,143,160,290,141,139,141,141,139],5.7,[139,143,165,292,141,139,141,141,139],7.81,[139,143,269,294,141,139,141,141,139],8.93,[139,143,272,296,141,139,141,141,139],9.89,[139,143,275,298,141,139,141,141,139],17.15,[139,143,278,300,141,139,141,141,139],22,[139,143,99,302,141,139,141,141,139],27.17,[139,150,139,304,141,139,139,141,139],6.27,[139,150,143,306,141,139,139,141,139],6.82,[139,150,150,308,141,139,139,141,139],8.58,[139,150,155,310,141,139,139,141,139],9.26,[139,150,160,312,141,139,139,141,139],11.12,[139,150,165,314,141,139,139,141,139],11.86,[139,150,269,316,141,139,139,141,139],12.59,[139,150,272,318,141,139,139,141,139],13.35,[139,150,275,320,141,139,139,141,139],16.4,[139,150,278,322,141,139,139,141,139],16.56,[139,150,99,324,141,139,139,141,139],17.32,[],[226],[92],[],[330],"Synthetic 3-D curve, 200 points per frame, true r=[0.02,0.25,-0.15], t=[40,120,-50]; noise std varied; 15 iterations; mean of 10 tries",{"slug":332,"group":333,"sourceId":5,"sourceLabel":6,"table":334,"selfRows":335,"metrics":336,"seqs":340,"entrants":361,"cells":363,"outcomes":419,"locators":420,"hardware":421,"wordings":422,"notes":423},"zhang1994icp-table-3","zhang1994icp:Table 3","Table 3",30,[337,338,339],{"label":108,"unit":109,"statistic":110,"alignment":44},{"label":112,"unit":109,"statistic":110,"alignment":44},{"label":234,"unit":115,"statistic":110,"alignment":44},[341,343,345,347,349,351,353,355,357,359],{"dataset":118,"sequence":342,"environment":120},"fraction of points 1",{"dataset":118,"sequence":344,"environment":120},"fraction of points 1\u002F2",{"dataset":118,"sequence":346,"environment":120},"fraction of points 1\u002F3",{"dataset":118,"sequence":348,"environment":120},"fraction of points 1\u002F4",{"dataset":118,"sequence":350,"environment":120},"fraction of points 1\u002F5",{"dataset":118,"sequence":352,"environment":120},"fraction of points 1\u002F6",{"dataset":118,"sequence":354,"environment":120},"fraction of points 1\u002F7",{"dataset":118,"sequence":356,"environment":120},"fraction of points 1\u002F8",{"dataset":118,"sequence":358,"environment":120},"fraction of points 1\u002F9",{"dataset":118,"sequence":360,"environment":120},"fraction of points 1\u002F10",[362],{"name":254,"methodId":5,"linkable":134,"proposed":134,"self":134},[364,365,367,369,371,373,375,377,379,381,383,384,386,388,389,391,393,395,397,398,400,401,403,405,407,409,411,413,415,417],[139,139,139,259,141,139,141,141,139],[139,139,143,366,141,139,141,141,139],3.44,[139,139,150,368,141,139,141,141,139],4.19,[139,139,155,370,141,139,141,141,139],4.88,[139,139,160,372,141,139,141,141,139],4.09,[139,139,165,374,141,139,141,141,139],7.52,[139,139,269,376,141,139,141,141,139],4.75,[139,139,272,378,141,139,141,141,139],6.09,[139,139,275,380,141,139,141,141,139],5.98,[139,139,278,382,141,139,141,141,139],4.9,[139,143,139,146,141,139,141,141,139],[139,143,143,385,141,139,141,141,139],5.14,[139,143,150,387,141,139,141,141,139],4.27,[139,143,155,376,141,139,141,141,139],[139,143,160,390,141,139,141,141,139],4.11,[139,143,165,392,141,139,141,141,139],6.67,[139,143,269,394,141,139,141,141,139],8.54,[139,143,272,396,141,139,141,141,139],7.45,[139,143,275,163,141,139,141,141,139],[139,143,278,399,141,139,141,141,139],7.34,[139,150,139,306,141,139,139,141,139],[139,150,143,402,141,139,139,141,139],3.53,[139,150,150,404,141,139,139,141,139],2.41,[139,150,155,406,141,139,139,141,139],1.85,[139,150,160,408,141,139,139,141,139],1.52,[139,150,165,410,141,139,139,141,139],1.28,[139,150,269,412,141,139,139,141,139],1.11,[139,150,272,414,141,139,139,141,139],1.01,[139,150,275,416,141,139,139,141,139],0.89,[139,150,278,418,141,139,139,141,139],0.83,[],[334],[92],[],[424],"Same synthetic curve, noise std 2 added to both curves; fraction of first-frame points varied from 1 to 1\u002F10; mean of 10 tries (iteration count not restated; the fraction-1 column equals the Table 2 std-2 column)",{"slug":426,"group":427,"sourceId":5,"sourceLabel":6,"table":428,"selfRows":269,"metrics":429,"seqs":434,"entrants":437,"cells":442,"outcomes":455,"locators":456,"hardware":458,"wordings":459,"notes":460},"zhang1994icp-text-sec-5-2","zhang1994icp:Text Sec.5.2","Text Sec.5.2",[430,431,432],{"label":108,"unit":109,"statistic":110,"alignment":44},{"label":112,"unit":109,"statistic":110,"alignment":44},{"label":433,"unit":115,"statistic":110,"alignment":44},"execution time (s)",[435],{"dataset":118,"sequence":436,"environment":120},"noise std 3",[438,440],{"name":439,"methodId":5,"linkable":134,"proposed":134,"self":134},"proposed with coarse-to-fine sampling",{"name":441,"methodId":5,"linkable":134,"proposed":134,"self":134},"proposed, all points, 15 iterations",[443,445,447,449,451,453],[139,139,139,444,141,139,141,141,139],4.56,[139,143,139,446,141,139,141,141,139],4.29,[139,150,139,448,141,139,139,141,139],3.39,[143,139,139,450,141,139,141,141,139],4.68,[143,143,139,452,141,139,141,141,139],4.14,[143,150,139,454,141,139,139,141,139],7.49,[],[457],"Sec. 5.2",[92],[],[461],"Noise std 3; coarse-to-fine (5 iterations with 1 of 5 points, then 10 with all) vs 15 iterations with all points; mean of 10 experiments",[463,469,474,479,485,491],{"group":464,"slug":465,"sourceLabel":6,"table":466,"selfRows":269,"datasets":467},"zhang1994icp:Text Sec.6.1","zhang1994icp-text-sec-6-1","Text Sec.6.1",[468],"rock scene (authors' data)",{"group":470,"slug":471,"sourceLabel":6,"table":472,"selfRows":269,"datasets":473},"zhang1994icp:Text Sec.7.2","zhang1994icp-text-sec-7-2","Text Sec.7.2",[468],{"group":475,"slug":476,"sourceLabel":6,"table":477,"selfRows":155,"datasets":478},"zhang1994icp:Text Sec.5.1","zhang1994icp-text-sec-5-1","Text Sec.5.1",[118],{"group":480,"slug":481,"sourceLabel":6,"table":482,"selfRows":155,"datasets":483},"zhang1994icp:Text Sec.5.3","zhang1994icp-text-sec-5-3","Text Sec.5.3",[484],"chair scene (authors' data)",{"group":486,"slug":487,"sourceLabel":6,"table":488,"selfRows":150,"datasets":489},"zhang1994icp:Text Sec.6.2","zhang1994icp-text-sec-6-2","Text Sec.6.2",[490],"head figure range images",{"group":492,"slug":493,"sourceLabel":6,"table":494,"selfRows":150,"datasets":495},"zhang1994icp:Text Sec.7.3","zhang1994icp-text-sec-7-3","Text Sec.7.3",[484],1790510662871]