• Laser & Optoelectronics Progress
  • Vol. 56, Issue 9, 091009 (2019)
Siyong Fu1,2 and Lushen Wu1,*
Author Affiliations
  • 1 School of Mechatronics Engineering, Nanchang University, Nanchang, Jiangxi 330031, China
  • 2 ZTE School of Communication and Information Engineering, Xinyu University, Xinyu, Jiangxi 338024, China
  • show less
    DOI: 10.3788/LOP56.091009 Cite this Article Set citation alerts
    Siyong Fu, Lushen Wu. Feature Extraction from 3D Point Clouds Based on Linear Intercept Ratio[J]. Laser & Optoelectronics Progress, 2019, 56(9): 091009 Copy Citation Text show less
    Schematic of linear intercept
    Fig. 1. Schematic of linear intercept
    Linear intercepts between two points for different cases. (a) dp01 for feature point of p0; (b) dp10for non-feature point of p1; (c) d'p01 for feature point of p0; (d) dp12, dp21 for non-feature points of p1 and p2
    Fig. 2. Linear intercepts between two points for different cases. (a) dp01 for feature point of p0; (b) dp10for non-feature point of p1; (c) d'p01 for feature point of p0; (d) dp12, dp21 for non-feature points of p1 and p2
    Phenomenon of misjudgment
    Fig. 3. Phenomenon of misjudgment
    Feature points of Fandisk model. (a) Original model; (b) model reduced by 60%; (c) model with 20 dB noise
    Fig. 4. Feature points of Fandisk model. (a) Original model; (b) model reduced by 60%; (c) model with 20 dB noise
    Feature points of Bunny model. (a) Original model; (b) model with 10 dB noise; (c) model reduced by 60%
    Fig. 5. Feature points of Bunny model. (a) Original model; (b) model with 10 dB noise; (c) model reduced by 60%
    Package diagram of model prototype. (a) Cross cuboid model; (b) workpiece model
    Fig. 6. Package diagram of model prototype. (a) Cross cuboid model; (b) workpiece model
    Feature points of cross cuboid model extracted under different δ values. (a) δ=10; (b) δ=9; (c) δ=8; (d) δ=7; (e) δ=6; (f) δ=5; (g) δ=4; (h) δ=3; (i) δ=2; (j) δ=1
    Fig. 7. Feature points of cross cuboid model extracted under different δ values. (a) δ=10; (b) δ=9; (c) δ=8; (d) δ=7; (e) δ=6; (f) δ=5; (g) δ=4; (h) δ=3; (i) δ=2; (j) δ=1
    Feature points of workpiece model extracted under different δ values. (a) δ=10; (b) δ=9; (c) δ=8; (d) δ=7; (e) δ=6; (f) δ=5; (g) δ=4; (h) δ=3; (i) δ=2; (j) δ=1
    Fig. 8. Feature points of workpiece model extracted under different δ values. (a) δ=10; (b) δ=9; (c) δ=8; (d) δ=7; (e) δ=6; (f) δ=5; (g) δ=4; (h) δ=3; (i) δ=2; (j) δ=1
    Extraction results of MSSV method under different intensities of noise. (a) 0 dB; (b) 5 dB; (c) 10 dB; (d) 15 dB
    Fig. 9. Extraction results of MSSV method under different intensities of noise. (a) 0 dB; (b) 5 dB; (c) 10 dB; (d) 15 dB
    Extraction results of NASD method under different intensities of noise. (a) 0 dB; (b) 5 dB; (c) 10 dB; (d) 15 dB
    Fig. 10. Extraction results of NASD method under different intensities of noise. (a) 0 dB; (b) 5 dB; (c) 10 dB; (d) 15 dB
    Extraction results of proposed method under different intensities of noise. (a) 0 dB; (b) 5 dB; (c) 10 dB; (d) 15 dB
    Fig. 11. Extraction results of proposed method under different intensities of noise. (a) 0 dB; (b) 5 dB; (c) 10 dB; (d) 15 dB
    Number of extracted feature points under different intensities of noise
    Fig. 12. Number of extracted feature points under different intensities of noise
    Extraction results of MSSV method. (a) Original model; (b) model reduced by 10%; (c) model reduced by 30%; (d)model reduced by 50%; (e) model reduced by 70%
    Fig. 13. Extraction results of MSSV method. (a) Original model; (b) model reduced by 10%; (c) model reduced by 30%; (d)model reduced by 50%; (e) model reduced by 70%
    Extraction results of NASD method. (a) Original model ; (b) model reduced by 10%; (c) model reduced by 30%; (d) model reduced by 50%; (e) model reduced by 70%
    Fig. 14. Extraction results of NASD method. (a) Original model ; (b) model reduced by 10%; (c) model reduced by 30%; (d) model reduced by 50%; (e) model reduced by 70%
    Extraction results of proposed method. (a) Original model; (b) model reduced by 10%; (c) model reduced by 30%; (d) model reduced by 50%; (e) model reduced by 70%
    Fig. 15. Extraction results of proposed method. (a) Original model; (b) model reduced by 10%; (c) model reduced by 30%; (d) model reduced by 50%; (e) model reduced by 70%
    Rate ofReduction /%Number of feature pointsComputation time /ms
    MSSVNASDProposed methodMSSVNASDProposed method
    0207562940539784424.51404.75390.42
    10186832335435810331.72305.41288.11
    30145311816427852211.31185.01161.57
    50103791297419894150.34130.2489.45
    70622777841193681.2162.2432.45
    Table 1. Number of feature points and computation time