摘要
针对轴孔零件精密装配过程中单目视觉定位系统的使用困难,设计了一种基于单目视觉的轴孔零件定位系统,提出了基于图像特征的零件倾斜角计算方法,可直接判定零件倾斜姿态。通过对椭圆快速检测算法进行优化,减小噪声影响,实现了高精度检测。最终通过该系统实现了轴孔零件的准确定位,提高了轴孔零件自动装配的成功率。实验证明,提出的方法能够在单目视觉条件下准确得到待装配零件的空间位置信息,满足同轴度要求15μm级的轴孔精密装配需求,具有较好的工程应用价值。
Aiming at the difficulty of using monocular vision positioning system in the process of precision assembly of shaft hole parts,a positioning system of shaft hole parts based on monocular vision is designed,and a method of calculating the inclination angle of parts based on image features is proposed,which can directly determine the inclination attitude of parts.By optimizing the fast ellipse detection algorithm,the influence of noise is reduced,and the highprecision detection is realized.Finally,through the system,the accurate positioning of the shaft hole parts is realized,and the success rate of the automatic assembly of the shaft hole parts is improved.The experimental results show that the proposed method can accurately obtain the spatial position information of the parts to be assembled under the condition of monocular vision,and meet the requirement of precision assembly of shaft hole with coaxiality of 15μm,which has good engineering application value.
作者
刘航铖
严小军
惠宏超
LIU Hangcheng;YAN Xiaojun;HUI Hongchao(Beijing Institute of Aerospace Control Devices,Beijing 100039,China)
出处
《机械与电子》
2021年第6期70-75,共6页
Machinery & Electronics
基金
装备预研航天科技联合基金(6141B061402)。
关键词
单目视觉
零件定位
椭圆拟合
二义性
monocular vision
part positioning
ellipse fitting
duality