空心微球具有广泛的应用, 耐压性能是其重要特性之一。但人工测量劳动强度大且精度较差。提出一种基于图像检测的自动测量方法, 可检测微球加压过程中的破损情况并记录压力值。首先, 利用梯度霍夫变换定位微球, 再通过Canny边缘检测提取微球轮廓信息, 并减少环境因素对检测的影响。最后Hu不变矩被用来将加压中的微球图像与加压前图像进行匹配, 当微球破损后, 匹配结果超过阈值, 判定微球失效。为证明方法的准确性和稳定性, 搭建了实验平台, 对不同直径的玻璃和塑料微球进行了加压实验。实验结果表明, 微球破损的成功识别率几乎为100%, 同时微球破损时的压力值也被准确地记录。

Pressure resistance is an essential characteristic of microspheres which has applications in various fields but the manual measurement of it is laborious and inaccurate. The method of automatic measurement using image detection to discriminate microspheres’ failures is proposed. Gradient Hough transform is used to localize microspheres. Contours are extracted via Canny edge detection and environment disturbances are also alleviated; finally Hu invariant moments are employed to implement image matching. A customized experiment platform was established and experiments involving glass and polymer microspheres with different diameters were conducted to examine the accuracy and robustness of the method. Nearly 100% success rate for discrimination of failed microspheres was achieved and the pressures at which microspheres failed were recorded with little delay.