基于显微视觉的微球圆度测量方法研究

Research on Measurement for Roundness of Microspheres Based on Micro-vision

利用显微视觉技术结合优化算法对微球的圆度进行自动测量。利用测量显微镜CCD相机采集微球图像,基于最大信息熵原理对图像增强,通过梯度非极大值抑制迭代算法分割图像,提取微球轮廓后进行最小二乘圆拟合,获得微球的圆度参数。该方法避免因滤波后微球目标边缘信息的丢失造成的检测精度降低。通过对0.5mm微球圆度测量实验证明:微球目标轮廓检测与提取的方法可以有效地抑制噪声干扰的影响,高效率实现目标边缘的准确定位;采用该方法获得的微球圆度值与标定微球圆度值之间的相对误差在0.17%之内,能够达到对微球圆度有效检测的目的。

The microscopic vision technology is used to automatically detect the roundness of microspheres with the optimization algorithm. The grabbed image of microsphere is acquired by the charge coupled device(CCD) of measuring microscope. The image is enhanced based on the principle of maximum information entropy. There is a algorithm is used to split image which proceed non-maximum suppression of the gradient and iterative. Then the roundness of microspheres by using least square circle fitting is obtained after trace pick-up. The method avoids the loss of some information of the image edge and the reduction of detection accuracy due to filtering processing. By measuring the roundness error of 0. 5 mm microspheres,it is proved that the method of edge detection and trace pick-up can effectively suppress the influence of noise interference and achieve the accurate location of the target edge. The relative error between the roundness of microspheres obtained by this method and the one of calibrated microspheres is within 0. 17%,which could achieved the purpose of effective detection of roundness.