多相机非共视场的非合作圆特征位姿测量方法

Non-cooperative Circle Characteristic Pose Measurement Using Multiple Cameras without Public Field of View

以大型光学模块转运时基于视觉的位姿测量为背景,利用目标底部同一平面的两个圆特征,针对大目标近距离对接时的环境限制,提出一种基于距离和角度约束的多相机非共视场位姿检测方法。对接时球头与锥孔正确对接,将两个相机分别固定于球头机构的内部后采集锥孔边缘的圆特征,利用非共视场成像的多相机标定获取两相机位置关系,融合多个相机的位姿信息,利用两圆共面和两相机位置关系约束剔除空间圆位姿解算时的虚假解。实验验证了该方法的精度,在1140mm的工作距离时,圆边缘特征的姿态角误差小于0.5°,圆心的计算误差小于1.0mm,实验结果表明该方法可准确计算位姿,结果可靠有效,在大目标近距离位姿测量时具有明显实用性。

Based on vision-based pose measurement during the trans-shipment of large optical modules,a non-public field of view(FOV)pose detection method for multi-cameras based on distance and angle constraints is proposed in this study.The method is developed by using two circular edges of the same plane at the bottom of the target of the installation module to eliminate the environmental constraints of close-range docking of large objects.While docking,the sphere of the transfer vehicle was correctly docked with the target cone hole.The two cameras were fixed in the hole of the sphere,and the circular features of the edge of the cone hole were collected.The position relationship between two cameras was obtained by multi-camera calibration in non-public FOV study on non-cooperative circle characteristic pose measurement using multiple cameras without public FOV imaging and the position and altitude information of multiple cameras were confluent.Finally,the false solutions in the pose calculation were eliminated using the constraints of two coplanar circles and two cameras'position relations.The experimental results show that the method can accurately calculate the pose.The calculation process is concise,and the results are reliable and effective.The error of the attitude angle of the circular edge feature is less than 0.5 degrees,and the calculation error of the center of the circle is less than 1.0mm at a working distance of 1140 mm.The proposed method has practical applications in measuring the position and altitude of large targets within close range.