Multi-Feature Vision Measurement Method for Aerospace Bolt Based on Contour Decomposition

Shuang Yan; Guofeng Wang; Jiefeng Li; Wei Tang; Zhizhuo Wang; Yanliang Sheng

doi:10.3788/LOP242077

Laser & Optoelectronics Progress, Volume. 62, Issue 10, 1012003(2025)

Multi-Feature Vision Measurement Method for Aerospace Bolt Based on Contour Decomposition

Shuang Yan¹, Guofeng Wang^1、*, Jiefeng Li¹, Wei Tang², Zhizhuo Wang¹, and Yanliang Sheng¹

Author Affiliations

¹School of Mechanical Engineering, Tianjin University, Tianjin 300354, China

²Aerospace Precision Products Inc., Ltd., Tianjin 300300, China

show less

Abstract Get PDF(in Chinese)

Figures & Tables(16)

Fig. 1. Framework of the bolt measurement algorithm

Download full size

Fig. 2. Principle of preliminary decomposition of contours. (a) Creating a contour binary tree; (b) process of contour stitching

Download full size

Fig. 3. Results of bolt contour decomposition. (a) Preprocessed contour; (b) preliminary decomposition of the contour

Download full size

Fig. 4. Effect of Pratt fitting error on contour decomposition. (a) Chamfered connection; (b) fillet connection

Download full size

Fig. 5. Optimization algorithm of contour decomposition

Download full size

Fig. 6. Results of bolt multi-feature recognition. (a) Optimization of contour decomposition; (b) multi-feature recognition of bolt

Download full size

Fig. 7. Bolt three-level feature tree

Download full size

Fig. 8. Vision measurement system for aerospace bolts

Download full size

Fig. 9. Comparison of bolt feature recognition algorithms

Download full size

Fig. 10. Measuring result distribution of bolt feature dimensions

Download full size

Table 1. Data structure of basic type contour
View table
Table 1. Data structure of basic type contour
struct BasicTypeContour
｛std::vector<cv::Point>bcontour；
_GEOMETRY_TYPE contourType；
double bpara［4］；
struct BasicGeoType * lcontour， * rcontour；
｝

Table 2. Data structure of bolt feature region
View table
Table 2. Data structure of bolt feature region
struct FeaRegionType
｛std::vector<BGT>fcontours；
_MP_FEATURE_TYPE featureType；
std::vector<fea_Data>cal_data；
struct FeaRegionType * lfeature， * rfeature；
｝

Table 3. Parameters of the hardware unit of the vision measurement system

View table

Table 3. Parameters of the hardware unit of the vision measurement system

Industrial camera		Lens		Light source
Parameter	Specifility	Parameter	Specifility	Parameter	Specifility
Model	MV-CE200	Model	MT027-110-C	Model	MT-L80-W
Sensor type	CMOS	Magnification	0.27	Type	Parallel light
Resolution	5472 pixel×3648 pixel	Focal length	110 mm	Beam diameter	80 mm
Pixel size	2.4 μm×2.4 μm	Mount type	C-mount	Color	White

Table 4. Aerospace bolts to be measured in experiments
View table
Table 4. Aerospace bolts to be measured in experiments
Bolt type Type Ⅰ Type Ⅱ Type Ⅲ Type Ⅳ
Aerospace bolt
Bolt length /mm 15.10 25.50 42.10 49.80

Table 5. Comparison of feature recognition algorithms

View table

Table 5. Comparison of feature recognition algorithms

Algorithm	Additional tools	Recall /%	Time /s	Accuracy of corner data /%	Recognize flaw
Template matching	Yes	81	2.1	0	No
Method in reference ［9］	Yes	89	1.6	0	No
Traditional contour recognition method	No	72	1.1	34	No
Ours	No	98	0.85	98	Yes

Table 6. Measuring error of bolt feature dimensions

View table

Table 6. Measuring error of bolt feature dimensions

Error type	Countersunk head			Fillet	Shank Ⅰ		Shank Ⅱ
Error type	H /mm	α /（′）	P-value /mm	R /mm	$d_{1}$ /mm	$L_{1}$ /mm	$d$ /mm	$L$ /mm
Mean absolute error	0.008	10	0.005	0.007	0.008	0.005	0.009	0.010
Max absolute error	0.016	13	0.014	0.013	0.010	0.013	0.010	0.015
Repeatability error	0.004	5	0.005	0.005	0.005	0.004	0.005	0.004

Tools

Get Citation

Copy Citation Text

Shuang Yan, Guofeng Wang, Jiefeng Li, Wei Tang, Zhizhuo Wang, Yanliang Sheng. Multi-Feature Vision Measurement Method for Aerospace Bolt Based on Contour Decomposition[J]. Laser & Optoelectronics Progress, 2025, 62(10): 1012003

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites