基于机器视觉的异形转子铜排截面积图像测量系统设计

Design of Image Measurement System for Cross-sectional Area of Spceial-shaped Rotor Copper Bar Based on Machine Vision

为提高异形转子铜排截面积的测量精度,解决传统方法不能稳定、重复且高精度完成测量的问题,根据异形转子铜排的特殊性设计其截面积的图像测量系统。图像采集单元设计包括专用的照明和成像设计,以摄取到高品质图像。设计的图像处理算法为:基于颜色差异信息提取异形转子铜排试样截面区域;通过与原始彩色图像进行匹配膨胀获得截面完整区域;采用面积微元法基于边界跟踪所获的截面边缘图像及所有边界点位置坐标计算铜排试样截面积。以标准仿真图像及多型号的实际试样为对象进行实验,验证了本系统的合理性和有效性。

To measure the cross-sectional area of spceial-shaped rotor copper bar with higher accuracy, an image measurement system based on the particularity of speeial-shaped rotor copper bar was developed, which solved problems of poor stability and low repeatability in traditional methods. Aiming to capture high quality image, the design for image acquisition unit consisted of appropriative imaging and lighting. The image processing algorithm worked as follows. By using color difference information, the spceial-shaped rotor copper sectional area of the test sample was extracted. Combining with original color image, the sectional area was completed by matching expansion algorithm. The processing of boundary tracking obtained sectional edge image and the coordinate positions of all boundary points. Based on the results of boundary tracking, the cross-sectional area of test sample was calculated by microelement method in area. Furthermore, taking a standard image and different types of test samples as objects, experiments verify rationality and validity of the system.