Research on Pose Detection of Shaft and Hole Parts Based on Monocular Endoscopic Imaging

doi:10.3969/j.issn.1004-132X.2026.09.003

Abstract

Abstract:

To ensure the quality of endoscopic imaging for detecting surface defects in inner holes of complex shaft parts and to avoid collision and scratching between the endoscope and the workpiece， a method was proposed herein for detecting the pose of shaft-hole parts based on monocular endoscopic imaging. Considering the characteristics of monocular endoscope imaging with many interference contours， a method was proposed for removing interference contours based on length constraint and minimum enclosing circle constraint. The improved Otsu method was adopted for image threshold segmentation， the Suzuki contour extraction algorithm was chosen for edge contour extraction， the proposed interference contour removal method was used for target contour selection， and the radial search was employed to extract the contour point set of the target inner hole. Considering that the contour point set was discrete and contained error points， and that incomplete target contours were caused by uneven illumination when the endoscopic light source was significantly offset in position and orientation， an ellipse detection algorithm was proposed based on contour restoration. Finally， considering the ambiguity problem in pose solving， a disambiguation method was proposed based on angle constraint and grayscale distribution，which leveraged the characteristics of endoscopic imaging：only circular features， rich grayscale information， the part being moved by the manipulator during pose detection， and the center of the inner hole being close to the optical axis. Experimental results verify the effectiveness of the proposed pose detection algorithm herein. The position detection accuracy in the X and Y axis directions is 0.062 mm， and the attitude detection accuracy around the X and Y axes is within 1.4°， which meets the requirements of enterprises.

Key words: pose detection, endoscopic imaging, ellipse detection, ambiguity removal, edge extraction, image segmentation

摘要：

为保障复杂轴类零件内孔表面缺陷检测的内窥成像质量并避免内窥镜与工件碰撞划伤，提出了一种轴孔类零件单目内窥成像位姿检测方法。针对单目内窥成像干扰轮廓多的特点，提出一种基于长度约束与最小包络圆约束的干扰轮廓去除方法。基于改进最大类间方差阈值分割方法对图像进行阈值分割，采用Suzuki轮廓提取算法进行边缘轮廓提取，采用所提出的干扰轮廓去除方法进行目标轮廓筛选并采用径向搜索提取目标内孔轮廓点集。针对轮廓点集离散且存在误差点，以及内窥光源位姿偏移较大时光照不均导致目标轮廓不完整的问题，提出了一种基于轮廓恢复的椭圆检测算法。针对位姿求解的二义性问题，基于成像仅有圆特征、灰度信息丰富且位姿检测时由机械臂移动工件、内孔中心靠近光轴的特点，提出一种基于角度约束与灰度分布的二义性去除方法。实验结果验证了提出的轴孔类零件单目内窥成像视觉位姿检测方法的有效性，该方法在X、Y轴方向上的位置检测精度为0.062 mm，绕X、Y轴旋转的姿态检测精度在1.4°以内，均达到了企业的要求。

关键词: 位姿检测, 内窥成像, 椭圆检测, 二义性去除, 边缘提取, 图像分割

CLC Number:

TP391.41

LI Rong, JIN Ziyuan, WANG Huajian, LIU Jian, YIN Shaohui. Research on Pose Detection of Shaft and Hole Parts Based on Monocular Endoscopic Imaging[J]. China Mechanical Engineering, 2026, 37(9): 2101-2114.

李蓉, 金子源, 王华建, 刘坚, 尹韶辉. 轴孔类零件单目内窥成像位姿检测研究[J]. 中国机械工程, 2026, 37(9): 2101-2114.

Figures/Tables 29

Fig.1 Appearance and profile of shaft part

Fig.2 Surface defect detection equipment & imaging scheme of the shaft-hole part based on machine vision

Fig.3 Four typical postures

Fig.4 Small end-face imaging with different posture offset

Fig.5 Comparison results of different segmentation algorithms

Fig.6 Comparison results of different threshold segmentation algorithms for inner hole with partially missing contour

Fig.7 Edge contour extraction result using Suzuki

Tab.1 Efficiency comparison of different contour extraction algorithms

轮廓提取方法	耗时/s
Canny算法	0.0055
形态学算法	0.0030
Suzuki算法	0.0020

Fig.8 Removal results of interference contour based on length constraint

Fig.9 Removal results of interference contour based on length constraint and minimum envelope circle constraint

Fig.10 Extraction of object contour points based on radial search

Fig.11 Ellipse detection flowchart of improved RED algorithm

Tab. 2 Parameter settings of improved RED algorithm

参数	数值
最大迭代次数 $T f$	100
用于初始椭圆拟合的6个边缘点每两个点应达到的最小距离 $T a$	30
边缘点到初始拟合椭圆边界的最大距离 $T d$	0.03
椭圆的完整度 $T r$	0.8

Tab. 2 Parameter settings of improved RED algorithm

参数	数值
最大迭代次数 $T f$	100
用于初始椭圆拟合的6个边缘点每两个点应达到的最小距离 $T a$	30
边缘点到初始拟合椭圆边界的最大距离 $T d$	0.03
椭圆的完整度 $T r$	0.8

Fig.12 Detection results of different ellipse detection methods

Fig.13 Flow chart of ellipse detection algorithm based on contour restoration

Fig.14 Contour restoration result

Fig. 15 Ellipse detection result based on contour restoration

Tab.3 Performance comparison of different ellipse detection methods

椭圆检测方法	每张图片耗时/s	未正确检出图片数	准确率/%
基于最小二乘法椭圆检测算法^［10］	0.698	30	93.724
基于弧段椭圆检测算法^［11］	12.817	19	96.025
无尺寸约束的本文改进的RED椭圆检测算法	0.729	4	99.163
本文提出的基于轮廓恢复的椭圆检测算法	0.719	0	100

Fig.16 Space circle projection diagram

Fig.17 Schematic of the region corresponding to the angle range

Fig.18 Gray-scale images of end-face with different postures of part

Tab. 4 Ambiguity judgment table based on angle constraint

真实法向量

虚假法向量

图18a

（ $-$ 0.033 192 13，

0.009 684 58，

$-$ 0.999 402 07）

（0.037 875 11，

$-$ 0.001 077 93，

$-$ 0.999 281 9）

图18b

（ $-$ 0.064 708 48，

0.034 254 27，

$-$ 0.997 316 13）

（0.073 214 31，

$-$ 0.048 404 26，

$-$ 0.996 140 9）

Tab. 4 Ambiguity judgment table based on angle constraint

真实法向量

虚假法向量

图18a

（ $-$ 0.033 192 13，

0.009 684 58，

$-$ 0.999 402 07）

（0.037 875 11，

$-$ 0.001 077 93，

$-$ 0.999 281 9）

图18b

（ $-$ 0.064 708 48，

0.034 254 27，

$-$ 0.997 316 13）

（0.073 214 31，

$-$ 0.048 404 26，

$-$ 0.996 140 9）

Fig.19 Gray-scale diagram

Fig.20 Theoretical and measured values of posture with different translation along XY axis

Fig.21 Theoretical and measured values of posture with rotation angles around the axis

Fig.22 Influence of radius disturbance

Fig.23 Influence of focal disturbance

Fig.24 Influence of principal disturbance

Fig.25 Influence of noise disturbance

References 25

[1]	孙露萍，张文昌，王志华，等. 基于2D激光位移传感器的舱段自动对接测量方法［J］. 中国机械工程， 2023， 34（9）： 1120-1125.
	SUN Luping， ZHANG Wenchang， WANG Zhihua， et al. Measurement Method of Cabin Automatic Docking Based on 2D Laser Displacement Sensor［J］. China Mechanical Engineering， 2023， 34（9）： 1120-1125.
[2]	陈栋，李世其，王峻峰，等. 一种以测量关键特征拟合舱段位姿的方法［J］. 中国机械工程， 2019， 30（18）： 2250-2256.
	CHEN Dong， LI Shiqi， WANG Junfeng， et al. A Method for Fitting Pose of Cabins through Measured Key Features［J］. China Mechanical Engineering， 2019， 30（18）： 2250-2256.
[3]	路恩会，郭耀村，李世雯，等. 基于全参考彩色图像质量算法的磨削表面粗糙度测量［J］. 仪器仪表学报， 2023， 44（7）： 252-259.
	LU Enhui， GUO Yaocun， LI Shiwen， et al. Measurement of Grinding Surface Roughness Based on the Full-reference Color Image Quality Algorithm［J］. Chinese Journal of Scientific Instrument， 2023， 44（7）： 252-259.
[4]	ZHANG Hang， LI Rong， ZOU Dexiang， et al. An Automatic Defect Detection Method for TO56 Semiconductor Laser Using Deep Convolutional Neural Network［J］. Computers & Industrial Engineering， 2023， 179： 109148.
[5]	SUO Xinyu， ZHANG Jie， LIU Jian， et al. Anomaly Detection in Annular Metal Turning Surfaces Based on a Priori Information and a Multi-scale Self-referencing Template［J］. Sensors， 2023， 23（15）： 6807.
[6]	LI Mengyuan， CHEN Ning， SUO Xinyu， et al. An Efficient Defect Detection Method for Nuclear-fuel Rod Grooves through Weakly Supervised Learning［J］. Measurement， 2023， 222： 113708.
[7]	刘坚，王媛媛. 一种面向球姿态检测的图像透视畸变校正方法［J］. 中国机械工程， 2019， 30（18）： 2242-2249.
	LIU Jian， WANG Yuanyuan. A Method of Image Perspective Distortion Correction in Ball Pose Detections［J］. China Mechanical Engineering， 2019， 30（18）： 2242-2249.
[8]	房运涛，王晓东，徐松，等. 万向支架微小螺纹副自动装配系统［J］. 中国机械工程， 2022， 33（6）： 698-706.
	FANG Yuntao， WANG Xiaodong， XU Song， et al. Miniature Thread Pairs Automatic Assembly System for Gimbals［J］. China Mechanical Engineering， 2022， 33（6）： 698-706.
[9]	黄鹏，郑淇，梁超. 图像分割方法综述［J］. 武汉大学学报（理学版）， 2020， 66（6）： 519-531.
	HUANG Peng， ZHENG Qi， LIANG Chao. Overview of Image Segmentation Methods［J］. Journal of Wuhan University （Natural Science Edition）， 2020， 66（6）： 519-531.
[10]	GANDER W， GOLUB G H， STREBEL R. Least-squares Fitting of Circles and Ellipses［J］. BIT Numerical Mathematics， 1994， 34（4）： 558-578.
[11]	LU Changsheng， XIA Siyu， SHAO Ming， et al. Arc-support Line Segments Revisited： an Efficient High-quality Ellipse Detection［J］. IEEE Transactions on Image Processing， 2020， 29： 768-781.
[12]	李良福，冯祖仁，贺凯良. 一种基于随机Hough变换的椭圆检测算法研究［J］. 模式识别与人工智能， 2005， 18（4）： 459-464.
	LI Liangfu， FENG Zuren， HE Kailiang. An Improved Algorithm for Ellipses Detection Based on Randomized Hough Transform［J］. Pattern Recognition and Artificial Intelligence， 2005， 18（4）： 459-464.
[13]	苗锡奎，朱枫，丁庆海，等. 基于星箭对接环部件的飞行器单目视觉位姿测量方法［J］. 光学学报， 2013， 33（4）： 412006.
	MIAO Xikui， ZHU Feng， DING Qinghai， et al. Monocular Vision Pose Measurement Based on Docking Ring Component［J］. Acta Optica Sinica， 2013， 33（4）： 412006.
[14]	MENG Cai， LI Zhaoxi， SUN Hongchao， et al. Satellite Pose Estimation via Single Perspective Circle and Line［J］. IEEE Transactions on Aerospace and Electronic Systems， 2018， 54（6）： 3084-3095.
[15]	ZHANG Dezhi， YANG Guocai， JI Junhong， et al. Pose Measurement and Motion Estimation of Non-cooperative Satellite Based on Spatial Circle Feature［J］. Advances in Space Research， 2023， 71（3）： 1721-1734.
[16]	HUANG Bin， SUN Yongrong， ZHU Yunfeng， et al. Vision Pose Estimation from Planar Dual Circles in a Single Image［J］. Optik， 2016， 127（10）： 4275-4280.
[17]	MA Xin， FENG Junbing， LI Yibin， et al. Active 6-D Position-pose Estimation of a Spatial Circle Using Monocular Eye-in-Hand System［J］. International Journal of Advanced Robotic Systems， 2018， 15： 1729881417753692.
[18]	王斌，吴丹，盖宇航. 面向机器人精密装配的高精度圆片位姿视觉检测［J］. 机械工程学报， 2023， 59（8）： 50-59.
	WANG Bin， WU Dan， GAI Yuhang. High-precision Wafer Pose Visual Detection for Robot Precise Assembly［J］. Journal of Mechanical Engineering， 2023， 59（8）： 50-59.
[19]	于长志，张连新，孙鹏飞，等. 基于法向约束的背靠视觉端面对接定位方法［J］. 计算机工程， 2022， 48（9）： 213-222.
	YU Changzhi， ZHANG Lianxin， SUN Pengfei， et al. End-face Butt-joint Positioning Method with Back Against Vision Based on Normal Constraint［J］. Computer Engineering， 2022， 48（9）： 213-222.
[20]	LI Shunzhi， LIANG Bin， GAO Xuehai， et al. Pose Measurement Method of Non-cooperative Circular Feature Based on Line Structured Light［C］∥2016 IEEE International Conference on Information and Automation （ICIA）.Changsha， 2016： 374-380.
[21]	MA Yanqin， LIU Xilong， XU De. Precision Pose Measurement of an Object with Flange Based on Shadow Distribution［J］. IEEE Transactions on Instrumentation and Measurement， 2020， 69（5）： 2003-2015.
[22]	余厚云，王慧青，张辉，等. 产品表面质量视觉检测中的相机位姿自动校准［J］. 东南大学学报（自然科学版）， 2020， 50（5）： 942-949.
	YU Houyun， WANG Huiqing， ZHANG Hui， et al. Automatic Calibration of Camera Pose in Visual Inspection of Product Surface Quality［J］. Journal of Southeast University （Natural Science Edition）， 2020， 50（5）： 942-949.
[23]	田光宝，王见，王博文. 单目相机非合作目标提取及位姿检测［J］. 红外与激光工程， 2021， 50（12）： 20210166.
	TIAN Guangbao， WANG Jian， WANG Bowen. Monocular Camera Non-cooperative Target Extraction and Pose Detection［J］. Infrared and Laser Engineering， 2021， 50（12）： 20210166.
[24]	LIU Yang， XIE Zongwu， LIU Hong. Fast and Robust Ellipse Detector Based on Edge Following Method［J］. IET Image Processing， 2019， 13（13）： 2409-2419.
[25]	闫蓓，王斌，李媛. 基于最小二乘法的椭圆拟合改进算法［J］. 北京航空航天大学学报， 2008， 34（3）： 295-298.
	YAN Bei， WANG Bin， LI Yuan. Optimal Ellipse Fitting Method Based on Least-square Principle［J］. Journal of Beijing University of Aeronautics and Astronautics， 2008， 34（3）： 295-298.