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：缸孔图象区域分割，二值化图象阈值的选取以及目标缺陷的确定。针对以上问题，采用直方图 均值适应法选取二值化阈值，对识别出的缺陷进行标定，并通过拟合方程得到缺陷的实际尺寸并最终确 定出目标缺陷。实验结果表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 machine servo system, the defects on the cylinder-bore's surface are detected automatically. The main problems solved in this paper are how to segment the image of the cylinder-bore; how to select threshold value; how to determine the object defects. To solve the problems, the value is chosen by the histogram average fitting method, the defects and the sizes are acquired by the regression formula. Experimental results show that the method has the features of high precision and high automation, which overcomes the weaknesses in traditional detecting method.
出处
《光学精密工程》
EI
CAS
CSCD
2000年第3期283-286,共4页
Optics and Precision Engineering
