Prediction and Compensation of Point Cloud Data Error in Line Laser Measurement

Shixiang Deng; Lü Yanming; Kang Wang; Kaixin Guo; Yin Zhang

doi:10.3788/LOP202259.1628006

Laser & Optoelectronics Progress, Volume. 59, Issue 16, 1628006(2022)

Prediction and Compensation of Point Cloud Data Error in Line Laser Measurement

Shixiang Deng¹, Lü Yanming^1,2、*, Kang Wang¹, Kaixin Guo¹, and Yin Zhang¹

Author Affiliations

¹School of Mechanical Engineering, Jiangnan University, Wuxi 214122, Jiangsu , China

²Jiangsu Provincial Key Laboratory of Advanced Food Manufacturing Equipment Technology, School of Mechanical Engineering, Jiangnan University, Wuxi 214122, Jiangsu , China

show less

Abstract Get PDF(in Chinese)

Figures & Tables(20)

Fig. 1. line laser on-machine measurement experimental platform

Download full size

Fig. 2. Schematic diagram of incidence angle detection device

Download full size

Fig. 3. Change value of incident angle error under standard reference distance of line laser sensor

Download full size

Fig. 4. Measurement error fitting diagram under different scanning positions and measurement results

Download full size

Fig. 5. Error results of RBF neural network training process

Download full size

Fig. 6. Comparison between predicted value and real value of data based on RBF neural network

Download full size

Fig. 7. Specific deviation diagram of predicted value and real value based on RBF neural network

Download full size

Fig. 8. Comparison between predicted value and real value based on SVR data

Download full size

Fig. 9. Specific deviation chart based on SVR predicted value and real value

Download full size

Fig. 10. Schematic diagram of line laser complex surface measurement

Download full size

Fig. 11. Schematic diagram of measurement of value points of laser aviation blade profile

Download full size

Fig. 12. On site measurement of aeroengine blades

Download full size

Fig. 13. Registration results and iterative change chart before ICP algorithm profile data compensation. (a) Registration results before profile data compensation; (b) change chart of RMS value with ICP iterations

Download full size

Fig. 14. Registration results and iterative change chart after ICP algorithm profile data compensation. (a) Registration results after profile data compensation; (b) change chart of RMS value with ICP iterations

Download full size

Table 1. Change of mean square error during RBF neural network training

View table

Table 1. Change of mean square error during RBF neural network training

NEWRB，neurons	MSE
0	0.066938700
4	0.002693020
8	0.000312165
12	0.000240473
16	0.000238059
20	0.000201481
24	0.000160629
28	0.000141652
32	0.000123842
36	0.000119751
40	0.000116904
44	0.000115682
48	0.000114716
52	0.000113483
56	0.000112347
60	0.000109987
64	0.000109732
68	0.000109956
72	0.000104075

Table 2. Relative error between predicted value and actual value based on RBF neural network
View table
Table 2. Relative error between predicted value and actual value based on RBF neural network
NO. Actual value /mm Predictive value /mm Relative error
1 0.0720 0.0780 0.08410
2 0.0873 0.1045 0.04482
3 0.1140 0.1306 0.14512
4 0.0939 0.0923 0.01686
5 0.0895 0.0909 0.01577

Table 3. Relative error between predicted value and actual value based on SVR
View table
Table 3. Relative error between predicted value and actual value based on SVR
No. Actual value /mm Predictive value /mm Relative error
1 0.0720 0.0747 0.03846
2 0.0873 0.0981 0.02837
3 0.1140 0.1126 0.01220
4 0.0939 0.0949 0.01028
5 0.0895 0.0903 0.00873

Table 4. Prediction effect comparison of two kinds of inclination error prediction models

View table

Table 4. Prediction effect comparison of two kinds of inclination error prediction models

No.	Actual value /mm	RBF neural network			SVR
No.	Actual value /mm	Predictive value /mm	Maximum error /mm	Relative error	Predictive value /mm	Maximum error /mm	Relative error
1	0.0720	0.0780	0.0165	0.08410	0.0747	0.0028	0.03846
2	0.0873	0.1045		0.04482	0.0981		0.02837
3	0.1140	0.1306		0.14512	0.1126		0.01220
4	0.0939	0.0923		0.01686	0.0949		0.01028
5	0.0895	0.0909		0.01577	0.0903		0.00873

Table 5. Comparison of prediction relative errors of two kinds of inclination error prediction models
View table
Table 5. Comparison of prediction relative errors of two kinds of inclination error prediction models
Error /% RBF neural network SVR
Average error 6.1334 1.9609
Maximum relative error 14.5115 3.8463

Table 6. Incidence angle compensation of line laser aviation blade

View table

Table 6. Incidence angle compensation of line laser aviation blade

Serial No.	Laser reading /mm	Measuring distance /mm	Coordinate converted value /mm	$Δ d$ /mm	$Δ x$ /mm	Inclination /（°）	Compen-sation value /mm	Correction value /mm
1	-0.5177	20.5177	25.4225	0.0016	0.01	8.8516	0.0098	25.4127
2	-0.4661	20.4661	25.4741	0.0032	0.01	17.9327	0.0153	25.4588
3	0.0322	19.9678	25.4724	0.0038	0.01	21.0091	0.0163	25.4561
4	1.5194	18.4806	25.4596	0.0006	0.01	3.1628	0.0081	25.4515
5	1.5044	18.4956	25.4446	0.0010	0.01	5.7821	0.0116	25.4330
$⋮$	$⋮$	$⋮$	$⋮$	$⋮$	$⋮$	$⋮$	$⋮$	$⋮$
75	2.2143	17.7857	-15.8455	0.0015	0.01	8.4867	0.0183	-15.8539
76	2.0085	17.9915	-16.0513	0.0006	0.01	3.5124	0.0099	-16.0588
77	1.8285	18.1715	-16.2313	0.0015	0.01	8.6117	0.0167	-16.2542
78	1.7883	18.2117	-16.2715	0.0023	0.01	13.1416	0.0584	-16.2927
79	1.3734	18.6266	-16.6864	0.0027	0.01	15.3326	0.0199	-16.6978
$⋮$	$⋮$	$⋮$	$⋮$	$⋮$	$⋮$	$⋮$	$⋮$	$⋮$
174	-0.2998	20.2998	-18.3596	0.0069	0.01	34.7126	0.0296	-18.3892
175	-0.3118	20.3118	-18.3716	0.0065	0.01	32.8651	0.0271	-18.3987
176	-0.3311	20.3311	-18.3909	0.0038	0.01	20.6513	0.0165	-18.4074
177	-0.3359	20.3359	-18.3957	0.0027	0.01	15.0172	0.0133	-18.4090
178	-0.3269	20.3269	-18.3867	0.0020	0.01	11.4123	0.0023	-18.3890

Tools

Get Citation

Copy Citation Text

Shixiang Deng, Lü Yanming, Kang Wang, Kaixin Guo, Yin Zhang. Prediction and Compensation of Point Cloud Data Error in Line Laser Measurement[J]. Laser & Optoelectronics Progress, 2022, 59(16): 1628006

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites

Paper Information

Category: Remote Sensing and Sensors

Received: May. 10, 2021

Accepted: Jul. 20, 2021

Published Online: Jul. 22, 2022

The Author Email: Yanming Lü (dsx654523115@126.com)

DOI:10.3788/LOP202259.1628006

Topics

laser devices and laser physics

Lasers and Laser Optics

Laser physics

laser manufacturing

Instrumentation, Measurement and Metrology

Table 1. Change of mean square error during RBF neural network training

Table 1. Change of mean square error during RBF neural network training

Table 2. Relative error between predicted value and actual value based on RBF neural network

Table 2. Relative error between predicted value and actual value based on RBF neural network

Table 3. Relative error between predicted value and actual value based on SVR

Table 3. Relative error between predicted value and actual value based on SVR

Table 4. Prediction effect comparison of two kinds of inclination error prediction models

Table 4. Prediction effect comparison of two kinds of inclination error prediction models

Table 5. Comparison of prediction relative errors of two kinds of inclination error prediction models

Table 5. Comparison of prediction relative errors of two kinds of inclination error prediction models

Table 6. Incidence angle compensation of line laser aviation blade

Table 6. Incidence angle compensation of line laser aviation blade