3 1,a) 1,b) 3D 3 3 Difference of Normals (DoN)[1] DoN, 1. 2010 Kinect[2] 3D 3 [3] 3 [4] 3 [5] 3 [6] [7] [1] [8] [9] [10] Difference of Normals (DoN) 48 8 [1] [6] DoN DoN 1 National Defense Academy a) em53035@nda.ac.jp b) iwak@nda.ac.jp 2. Difference of Normals(DoN) Ioannou DoN r ds r dl (r ds < r dl ) ( 2) 2.1 Jutzi (Principal Component Analysis=PCA) [11] 1 3 {λ 0 (n), λ 1 (n), λ 2 (n)} (λ 0 (n) λ 1 (n) λ 2 (n)) λ 2 (n) [12] 1 1
[6] DoN DoN 1 3.1 DDoN(Donuts DoN) DoN 4(2) DoN DDoN 3.2 RDoN(Ring DoN) 4(1) DoN 4(3) DoN RDoN 2 DoN 2.2 DoN PCA DoN DoN 2 DoN 2 0 2 PCA 0 DoN 3. DoN 4(1) DoN DoN 3.3 DoN PCA PCA 1 2 3(a) 3(a) DoN 0 1 χ χ = λ 0 (λ 0 + λ 1 + λ 2 ) (1) DoN 3 1 3 1 3(a) 3(b) 3(b) 2
(a) (b) 3 DoN χ s χ l DoN 3 DoN 4 3.4 DoC(Difference of Curvatures) DoC DoC χ χ 2 χ λ 0 λ 2 l,s χ = χ l χ s = λ 0s (λ 0s + λ 1s + λ 2s ) λ 0l (λ 0l + λ 1l + λ 2l ) (2) 3.5 DDoC(Donuts DoC) DDoC DDoN DoC 3 DDoN 3.6 RDoC(Ring DoC) RDoC RDoN DDoC 3 RDoC DDoC 3 DDoC (a) Front View (b) Side View 5 1 [mm] [mm] 7351 100x100x100 2.1 DDoC 4. 4.1 5 1 3D 10cm DoN [1] 5 (r ds, r dl ) = (1.5cm, 3.0cm) 3.7 4 DDoN 4.2 DoN n DoC c n c 3
6 4.3 DoN 6(a) DoN 2 DoN 4.4 DDoN 6(b) DDoN 2 7 9 7(a) 8(a) 9(a) DDoN DoN/DDoN 7(b) 7(c) 8(b) 8(c) 9(b) 9(c) DoN DDoN 9(b) 9(c) 2 DDoN DoN DDoN DoN 3 2 2 (b) DoN 7 (a) (c) DDoN 4.5 RDoN 6(c) RDoN 2 7 8 DoN DDoN (RDoN ) 2 3 2 9 DoN 2 RDoN (a) 4.6 DoC 6(d) DoC 2 3 2 (b) DoN 8 (c) DDoN 4
(a) DoN (b) DDoN (c) RDoN (d) DoC (e) DDoC (f) RDoC 6 DoN (a) 4.7 DDoC 6(e) DDoC 3 2 DDoC 3 DoN 3 DDoC (b) DoN (c) DDoN 9 4.8 RDoC 6(f) RDoC 2 3 DDoC 5
RDoC 3 DDoC DoN 3 2 DDoC 5. DoN DoN DoN 5 Library(PCL), Robotics and Automation, 4(2011). pp.1 [1] Y. Ioannou, B. Taati, R. Harrap, M. Greenspan: Difference of Normals as a Multi- Scale Operator in Unorganized Point Cloud, 3D Imaging, Modeling, Processing, Visualization and Transmission, pp.501 508(2012). [2] Microsoft: Kinect,http://www.xbox.com/ja- JP/kinect( 2015/7/2) [3] Microsoft: Kinect Fusion,https://msdn.microsoft.com/enus/library/dn188670.aspx( 2015/7/2) [4],,,, :, Vol. 76 No. 10,pp.1121 1124 (2010). [5] :, 2013 10-3,http://www.kensetsunews.com/?p=20090( 2015/7/2) [6],, : CCDoN: 6, Vol.80 No. 12,pp.1138 1143 (2014) [7],, : ND-PCA 3,,IS1-17(2015) [8] Kazuma Uenishi, Munetoshi Iwakiri: Virtual Keypoint Detector for 3D Registration, [9] K. Uenishi and M. Iwakiri: irtual Feature Point Extraction from Polyhe- dral Structur, Proceedings of IEEE International Symposium on Intelligent Signal Processing and Communication Systems, pp. 519 524 (2013). [10] A. Golovinskiy and T. Funkhouser: Min-cut based segmentation of point clouds, 2009 IEEE 12th International Conference on Computer Vision Workshops ICCV Workshops, 150:39 46(2009). [11] B. Jutzi, H. Gross: Nearest neighbour classification on laser point clouds to gain object structures from buildings The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences 38, 2009. [12] M. Pauly, M. Gross, L. Kobbelt: Efficient Simplification of Point-Sampled Surfaces,Proceedings of the conference on Visualization 02, pp.163 170(2002). [13] R. B. Rusu, S. Cousins: 3D is here: Point Cloud 6