Matching point pair optimization registration method for point cloud model

Table 1. Basic Information of spline matching point pairs
View table
View in Article
Table 1. Basic Information of spline matching point pairs
Standard（2 000） With noise（2 000） With noise and after iteration（2 000）
Count 167 127 941
Valid rate 0.083 5 0.063 5 0.470 5

Table 2. Basic information of registration experiment
View table
View in Article
Table 2. Basic information of registration experiment
Group Source Target $ξ$
1 Top3 Bun315 45%
2 Bun270 Bun315 60%
3 Bun090 Top3 75%
4 Bun000 Bun045 90%

Table 3. RMSE and registration time for different algorithms

View table

View in Article

Table 3. RMSE and registration time for different algorithms

Algorithm		Registration data
		Source	Target	Source	Target	Source	Target	Source	Bun000
		Top3	Bun315	Bun270	Bun315	Bun090	Top3	Target	Bun045
ICP	RMSE	$1.78 \times 10^{- 4}$		$0.97 \times 10^{- 4}$		$2.83 \times 10^{- 4}$		$1.99 \times 10^{- 4}$
ICP	Time/s	$4.28$		$4.34$		$3.44$		$4.08$
Tr-ICP	RMSE	$0.90 \times 10^{- 4}$		$0.51 \times 10^{- 4}$		$0.36 \times 10^{- 4}$		$0.53 \times 10^{- 4}$
Tr-ICP	Time/s	$5.13$		$4.82$		$3.81$		$4.16$
AA-ICP	RMSE	$0.81 \times 10^{- 4}$		$0.45 \times 10^{- 4}$		$0.34 \times 10^{- 4}$		$0.50 \times 10^{- 4}$
AA-ICP	Time/s	$3.15$		$3.01$		$3.13$		$3.65$
Ours	RMSE	$0.59 \times 10^{- 4}$		$0.39 \times 10^{- 4}$		$0.23 \times 10^{- 4}$		$0.42 \times 10^{- 4}$
Ours	Time/s	$3.47$		$3.17$		$3.29$		$3.83$

Table 4. RMSE for different algorithms

View table

View in Article

Table 4. RMSE for different algorithms

Algorithm	Number of source point clouds（Number of target point clouds）
Algorithm	20% 7 601（10 125）	25% 8 107（10 125）	30% 8 687（10 125）	35% 9 355（10 125）	40% 10 136（10 125）
ICP	-	-	-	-	-
Tr-ICP	-	-	$1.72 \times 10^{- 4}$	$1.53 \times 10^{- 4}$	$0.95 \times 10^{- 4}$
AA-ICP	-	-	$1.53 \times 10^{- 4}$	$1.21 \times 10^{- 4}$	$0.84 \times 10^{- 4}$
Ours	-	$1.08 \times 10^{- 4}$	$0.72 \times 10^{- 4}$	$0.62 \times 10^{- 4}$	$0.58 \times 10^{- 4}$

Table 5. RMSE and registration time of each algorithm under different levels of noise

View table

View in Article

Table 5. RMSE and registration time of each algorithm under different levels of noise

Algorithm		Number of noise points
Algorithm		N=1 000	N=2 000	N=3 000	N=4 000	N=5 000
ICP	RMSE	$2.78 \times 10^{- 4}$	$2.25 \times 10^{- 4}$	$2.46 \times 10^{- 4}$	$1.74 \times 10^{- 4}$	$2.59 \times 10^{- 4}$
ICP	Time/s	$4.21$	$3.65$	$4.14$	$3.99$	$4.32$
Tr-ICP	RMSE	$1.09 \times 10^{- 4}$	$1.17 \times 10^{- 4}$	$1.05 \times 10^{- 4}$	$1.19 \times 10^{- 4}$	$1.21 \times 10^{- 4}$
Tr-ICP	Time/s	$5.17$	$5.31$	$5.22$	$4.98$	$5.45$
AA-ICP	RMSE	$0.89 \times 10^{- 4}$	$0.94 \times 10^{- 4}$	$0.74 \times 10^{- 4}$	$0.96 \times 10^{- 4}$	$1.08 \times 10^{- 4}$
AA-ICP	Time/s	$3.12$	$3.27$	$3.19$	$3.20$	$3.49$
Ours	RMSE	$0.64 \times 10^{- 4}$	$0.69 \times 10^{- 4}$	$0.68 \times 10^{- 4}$	$0.72 \times 10^{- 4}$	$0.74 \times 10^{- 4}$
Ours	Time/s	$3.47$	$3.57$	$3.56$	$3.32$	$3.78$

Table 6. RMSE and time of different algorithms with gaussian noise
View table
View in Article
Table 6. RMSE and time of different algorithms with gaussian noise
Data Algorithm RMSE/ $10^{- 4}$ Time/s
Source：
Bun270
Target：
Bun315
ICP 5.95 6.38
Tr-ICP 1.30 6.96
AA-ICP 1.38 5.54
Ours 0.97 5.87

Table 7. Crankshaft registration RMSE and time
View table
View in Article
Table 7. Crankshaft registration RMSE and time
Data Algorithm RMSE/ $10^{- 4}$ Time/s
Source：
（143 610）
Target：
（129 136）
ICP 1.30 7.71
Tr-ICP 0.41 6.34
AA-ICP 0.38 4.29
Ours 0.27 5.32

Tools

Get Citation

Copy Citation Text

Yongwei YU, Kang WANG, Liuqing DU, Bing QU. Matching point pair optimization registration method for point cloud model[J]. Optics and Precision Engineering, 2023, 31(4): 503

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites