Aircraft Wing-Fuselage Fork-Ear Docking Assembly Deviation Detection Method Based on Binocular Vision

针对飞机翼身对接装配中叉耳位姿偏差测量精度不足的问题，本文提出了基于双目视觉的飞机翼身叉耳对接装配偏差检测方法。首先，搭建了飞机翼身叉耳对接装配偏差检测系统，然后，构建了基于双目视觉的叉耳装配偏差检测模型，采用RANSANC算法拟合叉耳孔边缘并剔除离群点，采用三角测量法并借助激光跟踪仪和参考板将相机坐标系中的孔心坐标转到全局坐标系。其次，采用三角测量法求出叉耳配合面上点的三维坐标并拟合边界直线。最终根据同轴度和间隙偏差数学模型计算装配偏差。实验结果表明，相比于传统的单目视觉检测，该方法有效降低了叉耳同轴度检测偏差和间隙检测偏差。To address the issue of insufficient measurement accuracy for pose deviations of forked lugs in aircraft wing-body docking assembly, this paper proposes a binocular vision-based method for detecting assembly deviations of aircraft wing-body forked lugs. First, an assembly deviation detection system for aircraft wing-body forked lugs was established. Then, a binocular vision-based detection model for forked lug assembly deviations was constructed. The RANSAC algorithm was employed to fit the edges of the lug holes and eliminate outliers. Triangulation was used in conjunction with a laser tracker and reference plate to transform the hole center coordinates in the camera coordinate system to the global coordinate system. Additionally, the triangulation method was applied to calculate the 3D coordinates of points on the mating surfaces of the lugs and fit the boundary lines. Finally, assembly deviations were computed according to the mathematical models of coaxiality and clearance deviation. Experimental results demonstrate that compared with traditional monocular vision detection, this method effectively reduces the detection deviations of lug coaxiality and clearance.