视觉导引下的轴与轴孔被动装配研究

提出了一种在视觉导引下的被动装配方法，该方法首先使用视觉进行粗定位，然后使用装配平台的扫描运动完成最后的装配，该方法克服了视觉装配误差大和被动装配时间长的缺点，集中了视觉装配速度快和被动装配精度高的优点，使得精确装配能够迅速完成。文中不仅给出了装配系统的模型，而且给出了空间点的矩阵变换，通过摄像机的特殊配置，使得矩阵变换非常简单。同时，对摄像机的最大成像误差进行了估计，为后面的装配平台运动提供依据。最后，给出了装配过程的实验结果，从装配过程所化的时间来看，该系统是可行的。

A New Passive Assembly Strategy for Peg-In-Hole Task Based on Vision

A new passive assembly strategy based on vision is introduced. According to this strategy, locate objective initially based on vision at first, and then make use of scan movement of assembly platform to finish final assembly. This strategy overcomes the faults that assembly based on vision is inaccurate and passive assembly is slow, and concentrates the advantages that assembly on vision is fast and passive assembly is accurate. By use of this strategy, precise assembly can be finished quickly. In this paper, we not only present the model of system, but also offer the transform of matrices; when placement of the camera is special, the transform of matrices is simple. At the same time, we estimate the most imaging error of camera, by which, we are able to control the move of assembly platform. In the end, an experiment of assembly course is provided to show its feasibility.