基于机器视觉的转速过程检测方法

Method of Detecting Rotation Process Based on Machine Vision

采用机器视觉技术对排种器排种盘的转速过程进行了检测。在排种盘上设置检测标志物，利用CCD摄像机采集排种盘在不同转速下的转速过程，并对获取的检测样本图像进行背景去除、二值化、图像平滑、特征量提取等处理，检测出转速序列{ni}和加速度序列{εi}，并对转速过程进行时序分析和误差估计。结果表明，在不同的设定转速下，机器视觉的转速检测相对误差范围为0．67％～3．40％。

Machine vision method was adopted to detect the rotation process. A detection-mark was mounted on the drill plate. The rotation process of the drill plat, under different rotational speeds, was shot by a CCD camera. A series of treatments were proposed including the deleting of backdrop, binary treatment, smooth image, the acquisition of feature values, and so on. Time intervals of rotation and angular acceleration were detected and analysed. The research shows that in different ranges of specified rotational speed, the relative error was from 0.67% to 3.4% . The detection has a broad application prospect in the estimate of machines＇ rotational stability and dynamic performance.