摘要
利用计算机视觉技术及相应的机械伺服系统,对汽车发动机缸孔表面缺陷进行检测,其主要解决的问题是:二值化图像阈值的选取,图像的分段检测,和缺陷的数据处理。针对以上的问题,采用最佳阈值分割法将图像二值化,并通过摄像机标定求取摄像机各参数,进而推导出标定缺陷实际面积的公式。实验结果证明此方法克服了传统检测方法的不足,是一种准确性和自动化程度较高的检测方法。
Using the computer vision technology and the corresponding machine servo system, the defects on the cylinder-bore's surface are detected.The main problems solved: how to select the threshold of the two value image, image's subsection detecting and the defects' data processing. Aiming at solving the above problems ,adopting the best threshold segment turns the image into two value, and the defects' geometry position and the area are obtained by spreading the cirque to rectangle. The experiment result shows that the method can conquer the deficiency of the traditional inspecting way.
出处
《计算机测量与控制》
CSCD
2005年第6期516-518,共3页
Computer Measurement &Control
