基于图像熵的直线电机动子位置精密测量方法

采用图像熵对栅栏条纹进行筛选,改进了图像相关测量方法,实现了对直线电机动子位置的精确实时测量。首先,分析了栅栏图像的Pnatt熵与像素相关性之间的关系,提出了采用图像熵表征栅栏图像筛选的特征值,从而筛选出非周期性更好的栅栏图像;其次,通过子区大小智能选择获得相邻两幅图像的计算区域,再采用粒子群优化算法对两计算区域的归一化互相关系数矩阵进行搜索,从而得到图像平移值,最后根据系统标定系数得到直线电机动子在拍摄间隔内的位移量。实验结果表明,基于图像熵筛选的栅栏图像在直线电机动子位置测量中具有更强的抗干扰性,测量精度达0.01像素。

Precise measurement of linear motor mover position using an improved digital image correlation based image entropy

Real-time and accurate measurement of the linear motor mover position is achieved by using an improved image correlation measurement method and the image entropy to filter image. Firstly, we analyzed the relationship between Pnatt entropy and pixel correlation of fence image. The image entropy was used as the eigenvalue of the fence image filter, so as to obtain a fence image with better aperiodic. Secondly, through adopting smart subset size selection method to get the region of interest, the normalized cross correlation matrix using particle swarm optimization algorithm was searched to obtain image translation value. Finally, the displacement of the linear motor was achieved based on the system calibration coefficient. Results show that the fence image based on image entropy screened is with a stronger anti-interference and 0.01 pixel accuracy in the linear motor mover position measurement.