目的 针对传送带输送产品过程中，机器人动态抓取产品的位姿数据计算问题，提出一整套基于智能相机和编码器检测的位姿数据计算方法。方法 首先搭建一套相机检测和机器人动态抓取系统模型，介绍抓取系统的构成及其工作过程；其次，为抓取系统创建各个坐标系，详细介绍各个坐标系间的变换关系、矩阵模型及需要标定的量；再次，设计标定量的测算过程和计算公式；最后，搭建以三菱工业机器人、欧姆龙智能相机及编码器等核心部件为主的实验样机，对上述位姿数据计算方法进行测试和验证。结果 实验结果表明，产品位姿数据的计算值与实测值之间的误差低于0.4 mm。结论 基于视觉和编码器检测的机器人动态抓取精度符合工程应用要求。

The work aims to propose a set of pose data calculation method based on intelligent camera and encoder test to solve the problem of calculating position and orientation data when the robot dynamically grasps the product during conveying. Firstly, a dynamically grasping system model of intelligent camera and encoder was built to introduce the composition and working principle of the grasping system. Secondly, each coordinate system of the grasping system was created to describe the transformation relationship, matrix model and data needing calibration among each coordinate system. Thirdly, the calculating process and formula for the calibration were designed. At last, the experimental prototype consists of the MITSUBISHI industrial robot, OMORON intelligent camera and encoder was developed, on which the above calculation method for the position and orientation data were tested and proved. The result shows that, the error between the calculated value and the measured value of the product’s position was less than 0.4 mm. The dynamic grasping precision based on vision and encoder test satisfies the requirements on engineering application.