基于SVD的直线电机动子位置的高精度测量方法

针对直线电机动子位置测量的高精度和实时性要求,提出了一种基于奇异值分解(SVD)相位相关算法的高精度图像测量方法。通过直线电机动子上的高速相机采集相邻栅栏图像,采用单行抽样栅栏条纹数据,再利用相位相关技术得到相邻栅栏图像的相位相关矩阵,进而利用SVD得到水平位移相位矢量。通过相位解缠算法和最小二乘法线性拟合得到二幅栅栏图像的亚像素位移量,将相邻栅栏条纹的像素位移值和动子实际位移进行标定,结果用于直线电机动子位置测量系统标定。并且进行傅里叶变换前采用Hanning窗抑制频谱泄露,提高匹配精度。实验结果表明,该直线电机动子位置的亚像素测量方法可以满足实时性和高精度测量要求。

An ultra precision measurement method based on the SVD for the linear motor mover position

Aiming at the requirement of high precision and real time measurement of the linear motor mover position, an image sub pixel measurement method based on the SVD phase correlation algorithm is proposed. The fence image of the motor stator is captured with a high speed micro camera installed on the linear motor mover in real time, according to single sampling to acquire pixel data of fence stripe and parallel computing. Then the phase correlation matrix of two adjacent fence images is achieved with the phase correlation algorithm, and the singular value decomposition is used to acquire horizontal displacement phase vector. The phase unwrapping algorithm and least squares linear fitting are used to get sub pixel displacement of two fences image, and the image pixel displacement and the mover displacement are made for the system calibration. As a result, the actual displacement of the linear motor mover is acquired. Meanwhile, the hanning window is added to inhibit the spectrum leakage before the Fourier transform, and the matching accuracy is improved. Experimental results show that the image sub pixel measurement method of the linear motor mover position can meet the requirement of real time and high precision measurement.